• Author / Uploaded
• sinanferda

Citation preview

The original French version, under the title Problemes de linguistique generale, was published in Paris. © Editions Gallimard, 1966 Copyright © 1971 by University of Miami Press Library of Congress Catalog Card No. 77-102692 SBN 87024-I32-X All rights reserved, including rights of reproduction and use in any fonn or by any means, including the making of copies by any photo process, or by electronic or mechanical device, printed or written or oral, or recording for sound or visual reproduction or for use in any knowledge or retrieval system or device, unless permission in writing is obtained from the copyright proprietors.

Designed by Mary Lipson Manufactured in the United States of America

Contents

Foreword vii Translator's Note Publisher's Note

ix x

Changes in Linguistics Recent Trends in General Linguistics 3 A Look at the Development of Linguistics 3 Saussure after Half a Century 29 I

2

17

Communication 4 The Nature of the Linguistic Sign 43 5 Animal Communication and Human Language 49 6 Categories of Thought and Language 55 7 Remarks on the Function of Language in Freudian Theory 65 Structures and Analyses 8 "Structure" in Linguistics 79 9 The Classification of Languages 85 10 The Levels of Linguistic Analysis 101 I I The Sublogical System of Prepositions in Latin I 13 12 Toward an Analysis of Case Functions: The Latin Genitive Syntactic Functions 13

The Nominal Sentence

131

14 Active and Middle Voice in the Verb 145 15 The Passive Construction of the Transitive Perfect 153 16 The Linguistic Functions of "To Be" and "To Have" 163

17 The Relative Clause, a Problem of General Syntax 181 Man and Language 18 19

Relationships of Person in the Verb 195 The Correlations of Tense in the French Verb

205

121

Contents

vi 20 21 22 23

The Nature of Pronouns 217 Subjectivity in Language 223 Analytical Philosophy and Language Delocutive Verbs 239

231

Lexicon and Culture 24 25 26 27 28

Semantic Problems in Reconstruction 249 Euphemisms Ancient and Modern 265 Gift and Exchange in the Indo-European Vocabulary 271 The Notion of "Rhythm" in its Linguistic Expression 281 Civilization: A Contribution to the History of the Word 289

Abbreviations Notes

297

Index 315

297

Foreword

THE STUDIES COLLECTED in this volume were chosen from among many others of a more technical nature which the author has published during the last few years. If they are here presented as "problems," it is because individually and as a group they make a contribution to the broad general problem of language as revealed in the principal topics taken up; we have considered the relations between the biological and the cultural, subjectivity and sociality, sign and object, and symbol and thought, as well as problems of intralinguistic analysis. Those who are discovering the importance of language in other areas will thus see how a linguist approaches some of the questions they have been led to ask themselves, and they will perceive how the configuration of language patterns all semiotic systems. To some readers, certain pages may seem difficult. They should be convinced that language is indeed a difficult subject and that the analysis of linguistic data is achieved by arduous paths. Like the other sciences, linguistics advances in direct proportion to the complexity which it recognizes in things; the stages of its development are the stages of this awareness. Moreover, one must bear in mind the truth that reflection on language is fruitful only if it deals first of all with real languages. The study of those empirical, historical organisms which actual languages are remains the only possible access to the understanding of the general mechanisms and functioning of language. In the first chapters we outline the trends in recent research in the theory of language and the prospects which have been opened up. We then go to the central problem of communication and its modalities: the nature of the linguistic sign, the distinctive characteristics of human langmJ1e, relations

viii

Foreword

between linguistic categories and categories of thought, and the role of language in the exploration of the unconscious. Notions of structure and function are the topics of the following essays, which deal successively with the variations of structure in languages and with the intralinguistic manifestations of some functions; among others, the connections between form and meaning are related to the levels of analysis. A separate series is devoted to syntactical phenomena: here we look for syntactic constants in very diversified linguistic types and set up specific models of certain sentence types to be recognized as universals: the nominal sentence and the relative clause. "Man and Language" is the title of the following part; here it is the mark of man upon language, defined by the linguistic forms of "subjectivity" and the categories of person, pronouns, and tense. In contrast, in the last chapters, it is the role of meaning and culture which is emphasized; there we study methods of semantic reconstruction as well as the origin of some terms of importance in modern culture. The unity and coherence of the whole will emerge from this survey. We have purposefully refrained from bringing in any later material, whether in the presentation or in the conclusions of the various chapters. Otherwise it would have been necessary to add an extended postscript to each of them, either with regard to the state of studies (for example, to indicate the most recent developments in theoretical research), or as the historian of our own research-the latter in order to give an account of the reception of these articles and to point out that "The Nature of the Linguistic Sign" (p. 43) provoked a lively controversy and brought forth a long series of articles, and that our pages on tense in the French verb (p. 205) were taken up and confirmed in the statistics of H. Yvon on the use of tense in modern writers, etc. But this would have been to initiate a new investigation each time. Other occasions will occur for coming back to these important questions and treating them anew. I should like to express my appreciation here to Messrs. P. Verstraeten and N. Ruwet for their part in the preparation of this volume.

E. B.

Translator's Note

The few changes or additions to the text necessitated by shifting from a French to an English frame of reference have been placed in square brackets. I wish to express my gratitude to Mr. Samuel N. Rosenberg for his kindness in reading over the translation and for his helpful criticism and advice.

M. E. M.

Publisher's Note

This translation was thoroughly checked and in some instances revised by the author. Certain editorial changes were also made in organization (such as placing all of the notes at the end of the volume) and in styling. Permissions to quote were generously granted by: University of Illinois Press, for material from G. J. Warnock, "Performative-Constative," in Philosophy and Ordinary Language, copyright 1963; North-Holland Publishing Company, for material in A. W. Groot's "Classification of the Uses of a Case Illustrated on the Genitive in Latin," in Lingua 6 (1956-1957); Diogene, for English translation of Emile Benveniste's article, "Animal Communication and Human Language," Diogenes, no. 1 (1952).

Changes in Linguistics

ONE

Recent Trends in General Linguistics

of the last decades, linguistics has developed so rapidly and extended its domain so far that even a cursory summary of the problems which it takes up would have to assume the proportions of a major work or else be condensed into a dry enumeration of individual efforts. It would take pages simply to sum up what has been learned and even then an essential development might be missing. The huge increase of productivity in linguistics is such that a thick volume is not enough to contain all the annual bibliography. The major countries now have their own publications, their collections, and also their methods. Efforts at description have been pursued and extended throughout the whole world; the recent edition of Les Langues du monde gives an idea of the work that has been accomplished and of the even greater amount that remains to be done. Linguistic atlases and dictionaries have multiplied. In all areas the accumulation of information produces works of ever increasing bulk: a description of the language of children in four volumes (W. F. Leopold), a description of French in seven volumes (Damourette and Pichon) are typical examples. It is possible today for an important review to be devoted exclusively to the languages of the American Indians. In Africa, in Australia, and in Oceania, research is being undertaken which is enriching considerably the inventory of linguistic forms. Parallel to this is the systematic exploration of the linguistic past of mankind. A whole group of ancient languages in Asia Minor has been brought within the province of Indo-European and this modifies the theory. The gradual reconstruction of proto-Chinese, common Malay-Polynesian, and of certain Amerindian prototypes will perhaps permit new genetic groupings, etc. But even if a much more detailed enumeration of research could be given, it would show that the work is proceeding very unevenly; here studies are continued which would have been the same in 1910, there even the term "linguistics" is rejected as being out of date, and elsewhere whole books are devoted just to the idea of the "phoneme." This multiplication of effort does not immediately reveal, but rather conceals, the profound changes which the

DURING THE COURSE

PROBLEMS IN GENERAL LINGUISTICS 4 methodology and spirit of linguistics have undergone in past decades, and the conflicts which divide it today. When one has opened his eyes to the importance of what is at stake and to the consequences which the current discussions might have for other disciplines, one is tempted to think that discussions about questions of methodology in linguistics may be only the prelude to a revision which will finally include all the sciences dealing with man. That is why we shall lay special stress, in nontechnical terms, on the problems which are today at the center of research in general linguistics, on the idea of their subject which linguists are developing and on the direction in which they are moving. Moreover, even as early as 1933, the volume published by the Journal de Psychologie under the title of Psychologie du langage demonstrated a remarkable burgeoning of theories and assertions of doctrine. There one could read the first expositions of principles which, like those concerning "phonology," are now taught everywhere. There also one could see the emergence of conflicts which have since led to reorganizations, such as the distinction between the synchronic and diachronic approaches and between phonetics and phonology, conflicts that disappeared when the terms were better defined. Certain convergences reconciled independent theories. When, for instance, Sapir brought to light the psychological reality of phonemes, he discovered on his own an idea which Trubetskoy and Jakobson had been working to establish. But one could not then foresee that in an ever broadening section of linguistics, research would, in appearance at least, run counter to the aims which linguistics had pursued up to then. I t has often been noted that the approach which characterized linguistics during the nineteenth century and at the beginning of the twentieth was exclusively historical. History as the necessary perspective and successivity as the principle of explanation, the splitting up of language into isolated elements and the investigation of the laws of evolution peculiar to each one of them-these were the dominant characteristics of linguistic doctrine. It was admitted, to be sure, that some principles of a completely different nature, like analogy, could disturb the regularity of the evolution. But in ordinary practice, the grammar of a language consisted of a presentation of the origin of each sound and of each form. This was the result both of the pervasiveness of evolutionism in all branches of learning and of the conditions under which linguistics came into existence. The novelty of the Saussurian point of view-one of those which has had the most profound effect-was to realize that language in itself does not admit of any historical dimension, that it consists of synchrony and structure, and that it only functions by virtue of its symbolic nature. It is not so much the historical point of view which is thus condemned as a certain way of "atomizing" language and of

Recent Trends in General Linguistics

5

making history mechanical. Time is not the agent of evolution; it is only its framework. The reason for the change that affects a certain element of a language lies, on the one hand, in the nature of the elements of which it is made up at a given moment, and on the other hand, in the structural relationships among those elements. The mere observation of the change and the formula of correspondence that sums it up make way for a comparative analysis of two successive states and of the different arrangements that characterize them. The legitimacy of diachrony, considered as a succession of synchronies, is thus reestablished. This brings out the prime importance of the idea of system and of the reestablislunent of interdependence among all the elements of a language. These views are today classic and were anticipated throughout the work of Meillet, and although they are not always applied, one would no longer find anyone to dispute them. If, from here on, the direction in which linguistics seems to be extending them today could be characterized in one word, it could be said that they mark the beginning of linguistics conceived of as a science, on account of its cohesiveness, its autonomy, and the aims which are assigned to it. This trend is indicated above all by the fact that certain types of problems have been abandoned. No one now seriously raises the question of the monogenesis or polygenesis of languages, or, in a general way, that of absolute origins. One no longer yields as easily as formerly to the temptation to erect the individual characteristics of a language or a linguistic type into universal qualities. The horizon of linguists has expanded. All types of languages have acquired equal rights to represent language in general. At no moment of the past and in no form of the present can one come upon anything "primordia1." The exploration of the most ancient attested languages shows them to be just as complete and no less complex than those of today; the analysis of "primitive" languages reveals a highly individualized and systematic organization in them. Far from constituting a norm, the Indo-European type appears instead to be rather the exception. All the more reason for turning away from research on a specific category recurring in most or all languages and believed to illustrate some universal tendency in "the mind of man"; we have seen how difficult it is to describe the whole system of a single language and how fallacious certain analogies of structure described by means of the same terms are. Great importance must be attached to this everbroadening experience of the linguistic diversity of the world. Several lessons have been extracted from this. First, it has become apparent that the conditions of evolution do not differ fundamentally according to the levels of culture, and that the methods and criteria that hold for languages of the written tradition could be applied to unwritten languages. From another

6

PROBLEMS IN GENERAL LINGUISTICS

point of view, it has been noticed that the description of certain linguistic types, notably of the Amerindian languages, raised problems which the traditional methods could not solve. The result has been a refurbishing of the procedures of description which, by a chain reaction, has been extended to languages which were thought to have been described for all time and which have taken on a new appearance. Still another consequence is the realization that the inventory of morphological categories, as diverse as it seems, is not without limits. One could thus imagine a sort of logical classification of these categories which would show their arrangement and laws of transformation. Finally, and here we touch upon questions whose range extends beyond linguistics, it can be seen that "mental categories" and "laws of thought" in large measure do nothing but reflect the organization and the distribution of linguistic categories. We imagine a universe which our language has first shaped. The varieties of philosophical or spiritual experience depend unconsciously on a classification which language brings about only for the reason that it is language and that it is symbolic. These are some of the topics that a cognizance of the diversity of linguistic types reveals, but to tell the truth, none of them has been thoroughly explored. To say that linguistics is becoming scientific is not only to stress the need for rigor which is common to all disciplines. It is a matter, first of all, of a change of attitude towards the subject, which could be defined as an effort at formalization. A double influence can be recognized at the beginning of this trend: that of Saussure in Europe and that of Bloomfield in America. The ways of their respective influences are as different as the works from which they proceeded. It is difficult to imagine a more marked contrast than that between these two works: the Cours de linguistique generale by Saussure (1916), a posthumous book put together from the notes of students, an ensemble of brilliant insights, each one of which calls for explication and certain of which still stir up controversy, setting language on the plane of a universal semiology and opening up views to which the philosophical thought of today is just awakening; and Language by Bloomfield (1933), which has become the vade mecum of American linguists, a "textbook" completely finished and matured, remarkable as much for its determined rejection of philosophy as for its technical rigor. Nevertheless, Bloomfield, although he does not refer to Saussure, would have certainly subscribed to the Saussurian principle that "The single true aim of linguistics is language envisaged in and of itself." This principle explains the trends which linguistics shows everywhere, even if it does not account for the reasons that make it wish to be autonomous and the ends it pursues thereby. Cutting across the differences of schools, the same preoccupations appear among linguists who try to systematize their procedures; these may be

Recent Trends in General Linguistics

7 formulated in three fundamental questions: (I) what is the task of the linguist, what does he reach, and what will he describe under the name of language? It is the very object of linguistics which is in question. (2) How will that aim be described? Tools must be forged that could grasp all the features of a language in all the known languages so as to describe them in identical terms. What then will be the principle of these procedures and of these definitions? This shows the importance which linguistic technique assumes. (3) For the naive speaker as for the linguist, the function of language is "to say something." What exactly is this "something" in regard to which language is articulated, and how is it defined with respect to language itself? The problem of meaning is raised. Simply listing these questions shows that the linguist wishes to break away from connections with or dependence upon previously established frames of reference or related disciplines. He rejects all a priori views of language in order to construct his notions directly from the object of his investigation. This attitude should put an end to the subordinate position, conscious or unconscious, in which linguistics found itself with regard to history on the one hand and to a certain psychology on the other. If the science of language has to choose models, it will be in the mathematical or deductive sciences, which completely rationalize their subject by reducing it to a set of objective characteristics subject to fixed definitions. That is to say that it will become more and more "formalistic," at least in the sense that language will consist of the totality of its observable "forms." Starting with native linguistic expression, one proceeds analytically to a thorough decomposition of each utterance into its elements, then by successive analyses to a decomposition of each element into increasingly simpler units. This operation will aim at b ringing out the distinctive units of a language, and this is already a radical change in methodology. While objectivity formerly used to mean total acceptance of data, which led to both the admission of the graphic norm in written languages and the minute recording of all the articulatory details for oral texts, today one strives to identify the elements insofar as they are distinctive at all levels of analysis. In order to recognize them, which is in no case an easy task, one is guided by the principle that there is nothing in a language but differences and that language avails itself of a system of discriminatory means. Only the features which are meaningful are retained by separating out, after having identified them, those which are simply variants. Great simplification results from this, and it now becomes possible to recog~ nize the internal organization and the laws of arrangement of these formal features. Each phoneme or morpheme becomes relative to each of the others, in that it is both different from and bound up with them, and each delimits the others, which in their turn delimit it, distinctiveness and solidarity being

8

PROBLEMS IN GENERAL LINGUISTICS

related conditions. These elements are organized in series and display in each language specific arrangements. It is a structure in which each part has its reason for being in the totality which it serves to compose. Structure is one of the essential terms in modern linguistics, one of those which still have a programmatic value. For those who use it on good grounds, and not simply for the sake of fashion, it can mean two very different things. By structure is meant, especially in Europe, the arrangement of a whole in parts and the demonstrable coherence of these reciprocally conditioned parts in the whole; for most of the American linguists, it is the distribution of the elements as it is observed, and the capacity of these elements for association or substitution. Because of this, the expression "structural linguistics" receives different interpretations; different enough, anyway, for the procedures derived from them not to have the same meaning. By the term structure, a "Bloomfieldian" will describe an arrangement of material which he will segmentalize into its constituent elements; he will define each of these elements by its position in the whole and by the variations and substitutions possible in this same position. He will reject as tainted with teleology the notion of balance and drift which Trubetskoy added to that of structure and which has since shown itself to be fruitful. It is indeed the only principle that makes the evolution of linguistic systems comprehensible. A state of a language is above all the result of a certain balance among the elements of a structure, a balance which, however, never leads to a complete symmetry, probably because dissymmetry is built into the very heart of language through the asymmetry of the organs of speech. The mutual dependence of all the elements causes every alteration at one point to have the repercussions throughout the system and sooner or later to produce a new arrangement. Consequently, diachronic analysis consists of setting up two successive structures and of stating their relationships, thereby showing what parts of the former system were altered or threatened and how the solution brought about by the later system was prepared. In this way the conflict between the synchronic and diachronic approaches so vigorously asserted by Saussure is solved. This concept of totally organized structure is completed by the notion of a hierarchy among the elements of the structure. A remarkable illustration of this is found in the analysis made by R. Jakobson of the acquisition and loss of the sounds of language by children and by aphasiacs respectively; the sounds that the child acquires last are the first the aphasiac loses, and those which the aphasiac loses last are the first to be articulated by children-the order of the disappearance is the reverse of that of acquisition. In any case, an analysis thus conceived is possible only if the linguist is able to observe in full, to control, or to vary at will the operation of the language described. Only living languages, written or not, offer a field large

Recent Trends in General Linguistics

9 enough and facts ascertainable enough for the investigation to be conducted with exhaustive rigor. The preponderance has been given to spoken languages. This condition has been imposed on certain linguists for empirical reasons. For others, in America, it was primarily the necessity of noting and analyzing the difficult and varied Indian languages, which was quite properly the point of departure for a revision in the methods of description and then in the general theory. But little by little the reorganization has been extended to the ancient languages. It has even become possible to reinterpret in the light of the new theories the data furnished by the comparative method. Works like those of J. Kurlyowicz on the reconstruction of the phases of Indo-European show how much can be expected of an analysis so oriented. A master of historical linguistics, J. Vendryes, also argues for a "static" linguistics that would be a comparative inventory of the resources which the different languages supply for the same requirements of expression. We can understand why the type of study that has predominated during these last years should be the systematic description, partial or total, of a particular language, with a more meticulous concern for technique than ever before, for the linguist feels constrained to justify his procedures from beginning to end. He puts forward an array of definitions that must account for the status he confers on each of the defined elements, and the operations are presented explicitly in such a way as to remain verifiable at all stages of the procedure. This results in an overhauling of the terminology. The terms employed are so specific that the informed linguist can recognize the approach of a particular study from its first lines, and certain discussions are intelligible to supporters of a method only when translated into their own terminology. One requires of a description that it be explicit and consistent and that the analysis be carried out without regard to meaning and use only formal criteria. It is especially in America that these principles are stated, and they have been extensively treated there. In a recent book, Methods of Structural Linguistics (1951), Z. S. Harris has produced a sort of codification of them. His work lists in detail, step by step, the processes for identifying phonemes and morphemes according to the formal conditions of their arrangemcntdistribution, environment, substitution, complementarity, segmentation, correlation, etc.-each of the operations being illustrated by particular problems treated with an almost mathematical array of graphic symbols. It would seem difficult to go further in this direction. Does it succeed at least in establishing a single and fixed method? The author is the first to agree that other procedures are possible and that certain of them would even be more economical, especially when meaning is introduced, so that one wonders finally if there is not some gratuitousness in this display of the demands of methodology. But, especially, it is to be observed that, in actual fact, all the work of the

10

PROBLEMS IN GENERAL LINGUISTICS

linguist deals with discourse implicitly assimilated to language. This point, which is fundamental, should be discussed along with the special concept of structure admitted by partisans of that method. Schemes of distribution, no matter how rigorously they are established, do not constitute a structure, any more than inventories of phonemes and morphemes, defined by means of segmentation in chains of discourse, represent a description of a language. \Vhat is given to us, in fact, is a method of transcribing and materially analyzing a language which is represented by a body of oral texts, the meaning of which the linguist supposedly does not know. Let us emphasize especially that feature which, even more than the special technique of the procedure, characterizes the method; it is the principle that linguistic analysis, in order to be scientific, should ignore meaning and apply itself solely to the definition and distribution of the elements. The conditions of rigor imposed on the procedure require that that elusive, subjective, and unclassifiable element which is meaning or sense be eliminated. All that can be done will be to make sure that a certain utterance corresponds to a certain objective situation, and if the recurrence of the situation elicits the same utterance, they are to be correlated. The relationship between form and sense is thus understood as the relationship between the linguistic expression and the situation, in terms of behaviorist theory, so that the expression might be both response and stimulus. Thus, meaning is practically reduced to a certain linguistic conditioning. As for the connection between the expression and the world, this is a problem to be left to the specialists in the physical universe. Bloomfield defines "the meaning of a linguistic form as the situation in which the speaker utters it and the response which it calls forth in the hearer" (Language, p. 139). And Harris stresses the difficulty of analyzing situations: "There are at present no methods of measuring social situations and of uniquely identifying social situations as composed of constituent parts, in such a way that we can divide the utterance which occurs in (or corresponds to) that social situation into segments which will correspond to the constituent parts of the situation. In general, we cannot at present rely on some natural or scientifically ascertainable subdivision of the meaning range of the local culture, because techniques for such complete cultural analysis into discrete elements do not exist today; on the contrary, language is one of our chief sources of knowledge about a people's culture (or 'world of meaning') and the distinctions or divisions which are made in it" (Methods, pp. 188-189). It is to be feared that if this method becomes general, linguistics may never be able to join any of the other sciences of man or of culture. The segmentation of the statement into discrete elements does not any more lead to an analysis of language than the segmentation of the physical universe leads to a theory of the physical world. This way of formalizing the parts of the state-

Recent Trends in General Linguistics

II

ment may lead to a new atomization of language, for an empirical language is the result of a process of symbolization at several levels, the analysis of which has not even been attempted. The linguistic "datum" is not, considered in this way, a primary datum with which all one has to do is break it down into its component parts; it is actually a complex, some of whose values come from particular qualities of the elements, others from the conditions of their arrangement, still others from the objective situation. One can then conceive of several types of description and several types of formalization, but all of them must necessarily assume that their object, language, is informed with meaning, which gives it its structure, and that that condition is essential to the functioning of language among other systems of signs. It is difficult to imagine the result of segmentalizing a culture into discrete elements. In a culture as in a language, there is a network of symbols whose relationship must be defined. Until now, the science of cultures has remained strongly and deliberately "substantial." Can formal structures of the type of those introduced by Claude Levi-Strauss into the systems of family relationship be recognized in the apparatus of culture? This is a problem of the future. In any case it can be seen how necessary an investigation of the properties of symbols would be for all sciences that operate with symbolic forms. The research begun by Peirce has not been followed up and this is a great pity. A better understanding of the complex processes of meaning in language and probably outside language could be expected, among other things, from progress in the analysis of symbols. And since this functioning is unconscious, just as the structure of behavior is unconscious, psychologists, sociologists, and linguists would do well tc join together in this research. The orientation which we have just characterized is not the only one to note. Other conceptions, equally systematic, have been advanced. In the psycho linguistics of G. Guillaume, linguistic structure is posited as inherent in an actual language, and the systematic structure discloses itself from the facts of usage which make it explicit. The theory which Louis Hjelmslev, in Denmark, wishes to put forth under the name of "glossematics" is a construction of a logical "model" of language and a body of definitions rather than an instrument for exploring the linguistic universe. The central idea here is roughly that of the Saussurian "sign," in which expression and content (corresponding to the Saussurian "signifiant" and "signifiP') are two correlated planes, each of which comprises a Clform" and a Clsubstance". The movement here is from linguistics toward logic. And in this connection can be seen what could be a convergence of sciences which are still largely unaware of one another. At the moment when linguists concerned with rigor seek to borrow the ways and even the apparatus of symbolic logic for their formal operations, it happens that logicians are becoming attentive to

12

PROBLEMS IN GENERAL LINGUISTICS

linguistic "meaning" and, following Russell and Wittgenstein, interest themselves increasingly in language. Their paths cross rather than meet, and logicians concerned with language do not always find linguists to converse with. As a matter of fact, those linguists who wish to insure linguistics a scientific status turn rather toward mathematics, seeking transcription devices rather than an axiomatic method, and they give in rather easily to the lure of certain recent techniques like the theories of cybernetics or of information. A more fruitful task would be to consider the means of applying in linguistics certain of the operations of symbolic logic. The logician scrutinizes the conditions of truth that must be satisfied by the statements that shape science. He will not accept "ordinary" language because it is equivocal, uncertain, and vague, and he strives to forge an entirely symbolic language. But the subject of the linguist is precisely this "ordinary language," which he takes as the datum and whose entire structure he explores. It would be to his interest to utilize tentatively, in the analysis of all orders of linguistic classes which he determines, the instruments developed by the theory of sets, in order to see if relationships amenable to logical symbolization can be stated among these classes. One would then at least have some idea of the type of logic underlying the organization of a language, one would see if there is a qualitative difference between the types of relations peculiar to ordinary language and those which characterize the language of scientific description, or, in other words, how the language of action and the language of intellect interreact. It is not enough to observe that one allows itself to be transcribed in a symbolic notation and the other does not, or does not immediately; the fact remains that both the one and the other proceed from the same source and that they contain the same basic elements. It is language itself that suggests the problem. These considerations appear to take us far away from those topics of research which linguistics was proposing a few decades ago. But these problems are those of all times, even if it is only now that they are being attacked. On the other hand, in the connections which linguists were then seeking with other fields, today we find difficulties which they hardly suspected. Meillet wrote in 1906: "It will be necessary to determine what social structure a given linguistic structure responds to and how, in a general way, changes in social structure are expressed in changes in linguistic structure." In spite of certain endeavors (Sommerfelt), this program has not been completed, for just as soon as it was attempted to compare language and society systematically, disparities emerged. It appeared that the correspondence between them was constantly disturbed by the major fact of diffusion, in language as well as in social structure, so that societies of the same culture may have heterogeneous languages, while languages very closely related may

Recent Trends in General Linguistics

13

be used for the expression of entirely dissimilar cultures. In pushing the consideration further, problems inherent in the analysis of language on the one hand and of culture on the other, and problems of "meaning," which are common to them both, were encountered-in short, the very problems mentioned above. This does not mean that the program of studies indicated by MeiIlet cannot be realized. The problem, instead, is to discover the common basis of language and society and the principles that govern these two structures by first defining the units that in both the one and the other would lend themselves to being compared; it is also to bring out their interdependence. There are naturally easier ways to approach the question, but actually they transform it; for example, the study of the imprint of culture on language. In practice, one is limited to the lexicon. It is then no longer a language that is dealt with, but the composition of its vocabulary. It is, however, a very rich subject, and, appearances notwithstanding, one which has been turned to hardly any account. Ample repositories of data that will foster many studies are now available, among them, J. Pokorny's Indogermanisches Etymologisches Worterbuch and C. D. Buck's Dictionary of Selected Synonyms in the Principal Indo-European Languages. The study of the variations in meaning during the course of history is another very promising field. Important works have been devoted to the "semantics" of vocabulary in its theoretical as well as in its social or historical aspects (Stern, Ullmann). The difficulty is to identify such constants as will allow the construction of a theory of lexical signification from the growing mass of empirical facts. These facts seem to offer a constant challenge to any predictability. From another point of view, the action of "beliefs" upon expression raises numerous questions, certain of which have been studied: the importance of linguistic taboo (Meillet, Havers), the modifications of linguistic forms to signal the attitude of the speaker toward the things he is speaking of (Sapir), the ceremonial hierarchy of expressions; and all of these bring to light the complex action of social behavior and psychological conditioning in the use of language. This leads to the problems of "style" in all its acceptations. In the course of these last years, studies of very different but equally important trends have dealt with the processes of style (Bally, Cressot, Marouzeau, Spitzer, Vossler). Since research along these lines makes use, consciously or not, of criteria which are at the same time aesthetic, linguistic, and psychological, it involves all at once the structure of the language, its powers of stimulation, and the reactions it provokes. Although the criteria remain only too often "subjective," care is taken to state precisely the method to deal with the affective contents, with the intention which produces them as well as with the language they spring from. Progress is being made in that direction with studies on word order, the quality of sounds, rhythm and prosody, as well as on the

PROBLEMS IN GENERAL LINGUISTICS

lexical and grammatical resources of the language. In this regard, psychology has been widely called upon, not only because of the affective notions constantly involved in the analysis, but also for the techniques used to objectify them-tests of evocation, research on color perception of sounds, vocal timbres, etc. It is a whole symbolism which one is slowly learning to decode. Thus one observes on all sides the effort to submit linguistics to rigorous methods and to banish from it mere approximation, subjective constructions, and philosophical a priori-ism. Linguistic studies are becoming increasingly difficult from the very fact of these demands and because linguists are discovering that language is a complex of specific qualities to be described by methods still to be created. So specific are the conditions peculiar to language that one could lay down as fact that there are not one but several structures of language, each one of which would give rise to a complete linguistics. To become aware of this will perhaps contribute to an understanding of the present conflicts. Language has, first of all, the eminently distinctive quality that it is always built up on two planes, those of the "significant" and the "signified." The study of that constituent quality of language and of the correlations of regularity or disharmony which it involves, of the tensions and transformations which result from it in each individual language, could alone serve as a basis for a linguistics. But language is also human; it is the point of interaction between the mental and the cultural life in man, and at the same time the instrument of that interaction. Another linguistics could be founded on the terms of this trinomial: language, culture, and personality. Language could also be considered as being contained in a body of articulated utterances, which would provide the basis of a strictly objective study. A language would then be the object of an exhaustive description which would proceed by means of a segmentation of observable data. On the other hand, this language realized in recordable utteraRces could be considered the contingent manifestation of a hidden substructure. It is, in that case, the seeking out and bringing to light of this latent mechanism which would be the object of linguistics. Language in general admits also of being constituted in a structure of game, like a pattern of "figures" produced by the intrinsic relationships of constant elements. Here linguistics would become the theory of possible combinations among these elements and the universal laws that govern them. Another possibility would be the study of language as a branch of a general semiotics which would cover both mental and social life. The linguist would then have to define the basic nature of linguistic symbols through a rigorous formalization and a distinct metalanguage. This enumeration is not exhaustive and cannot be. Other conceptions will perhaps appear. We only wish to show that behind the discussions and asser-

Recent Trends in General Linguistics tions of principle of which a glimpse has just been given, there is often, without all linguists seeing it clearly, a preliminary choice which determines the position of the object of investigation and the nature of the method. These diverse theories will probably coexist-although at one point or another of their development they must necessarily meet-until the moment when the status of linguistics as a science will impose itself, not as a science of empirical facts but as one of relations and deductions recapturing unity of plan in the infinite diversity of linguistic phenomena.

From Journal de Psychologie, Paris, January-June, 1954, pp. 139 ff.

TWO

A Look at the Development of Linguistics

DURING THE COURSE of these last years, extensive changes, whose scope extends even beyond the already very broad horizon of linguistics, have taken place in studies dealing with language and languages. These changes may not be understood all at once-they elude one even as they appear. In the long run, they have made difficult the approach to new works, which bristle with an increasingly technical terminology. It is a fact that great difficulty is experienced in reading linguists and even more in comprehending their concerns. What are they aiming at and what are they doing with what is the property. of ~ll men and never ceases to attract their curiosity-language? One has the impression that for the linguists of today, the facts of language have been transformed into abstractions, becoming the inhuman materials of algebraic constructions or serving as evidence in arid discussions of methodology; that linguistics is getting away from the realities of language and isolating itself from the other social sciences. Actually it is just the opposite. We observe at the same time that these new methods of linguistics are taken as examples and even as models for other sciences, that the problems of language are today of interest in very diverse and increasingly numerous fields of specialization, and that there is a trend in the research done in the social sciences toward working with the same mind that inspires linguisitics. It may then be useful to set forth, as simply as can be done with this difficult subject, how and why linguistics has thus been transformed, starting from its origins. Let us begin by observing that linguistics has a double scope: it is the science of language and the science of languages. This distinction, which is not always made, is necessary; language, the human faculty, the universal and immutable characteristic of man, is something distinct from particular languages, always individual and variable, in which it is realized. It is with languages that the linguist deals, and linguistics is primarily the theory of languages. But, seen from the point of view we choose here, these different paths will often intertwine and finally mingle, for the infinitely diverse

PROBLEMS IN GENERAL LINGUISTICS

problems of particular languages have in common that, when stated to a certain degree of generality, they always have a bearing on language in general. Everyone knows that western linguistics originated in Greek philosophy. This appears from all the evidence. Our linguistic terminology is made up in large part of Greek terms adopted directly or through their Latin translations. But the interest which the Greek thinkers very early took in language was exclusively philosophical. They argued about its original condition-was language natural or conventional?-instead of studying its functioning. The categories they established (noun, verb, grammatical gender, etc.) always rest on logical or philosophical grounds. For centuries, from the pre-Socratics to the Stoics and Alexandrians, then during the Aristotelian renaissance, which extended Greek thought until the end of the Latin Middle Ages, language remained an object of speculation, not of observation. No one was concerned with studying and describing a language for itself, or with inquiring whether the categories of Greek or Latin grammar had general validity. This attitude hardly changed until the eighteenth century. A new phase developed at the beginning of the nineteenth century with the discovery of Sanskrit. It was forthwith discovered that there was a relationship among the languages which have since been called Indo-European. Linguistics was worked out within the framework of comparative grammar with methods which became more and more rigorous as chance findings or decipherings gratified this new science with confirmations of its basic principles and expansions of its range. The work accomplished in a century was considerable and admirable. The methodology tested in the Indo-European field set the pattern. Revised today, it is having new successes. But it should be stated that up until the first decades of our century, linguistics consisted of what was mainly a genetics of languages. It took as its task the study of the evolution of linguistic forms. It set itself up as a historical science, its focus being, everywhere and always, a phase of the history of languages. However, in the midst of these successes, some people were disturbed. Whatis the nature of a linguistic phenomenon? What is the reality of language? Is it true that it consists of nothing but change? But how does it stay the same and change at the same time? How does it function and what is the relationship of sound to sense? Historical linguistics gave no answers to these questions, having never had to raise them. At the same time, difficulties of a different sort but equally formidable were in the offing. Linguists were beginning to turn to languages which were not written and had no history, especially the Indian languages in America, and they discovered that the traditional frameworks used with the Indo-European languages did not apply to them. They came

A Look at the Development of Linguistics

upon completely different categories which eluded a historical descrip tion and called for a new array of definitions and a new method of analysis. Little by little, through many theoretical discussions and under the inspiration of the Cours de linguistique gentrale of Ferdinand de Saussure (1916), a new notion of language took shape. Linguists came to realize what was incumbent upon them: to study and describe the present linguistic reality by an adequate technique and not to mix any theoretical or historical presupposition with the description, which should be synchronic, and to analyze the language into its own formal elements. Linguistics thus entered into its third phase-the present one. It took for its subject not the philosophy of language nor the evolution of linguistic forms, but brought to the fore the intrinsic reality of language, and aimed at making itself a formal, rigorous, and systematic science. With that, both the historical considerations and the frameworks set up for the Indo-European languages were again called in question. In becoming descriptive, linguistics granted equal importance to all types of languages, written or not, and it had to adapt its methods to them. The problem was indeed to know what a language consisted of and how it functioned. When, following the lead of F. de Saussure, linguists began to envisage language in and of itself, they recognized what was to become the basic principle of modern linguistics: that language forms a system. This holds true for any language, no matter in what culture it is in use, at whatever historical stage we view it. From base to summit, from sounds to the most complex forms of expression, language is a systematic arrangement of parts. It is made up of formal elements put together in variable combinations. according to certain principles of structure. Here is the second key term in linguisticsstructure. By this is meant primarily the structure of the linguistic system gradually revealed. starting from the principle that a language always has a limited number of basic elements. but that these elements. few in number in themselves, yield a large number of combinations. They are. indeed, only grasped within these combinations. Now, methodical analysis leads to the realization that a language actually uses only a small part of the theoretically numerous combinations that would result from these minimal elements being freely assembled. That restriction makes certain specific configurations take shape. which will vary according to the linguistic systems under consideration. Here is what is meant primarily by structure: particular types of relationships articulating the units of a certain level. Each one of the units of a system is thus defined by the relations which it maintains with the other units and by the oppositions into which it enters; as Saussure says. it is a relating and opposing entity. Thus the idea that the data of a language have value in themselves and are objective "facts," absolute

PROBLEMS IN GENERAL LINGUISTICS

problems of particular languages have in common that, when stated to a certain degree of generality, they always have a bearing on language in general. Everyone knows that western linguistics originated in Greek philosophy. This appears from all the evidence. Our linguistic terminology is made up in large part of Greek terms adopted directly or through their Latin translations. But the interest which the Greek thinkers very early took in language was exclusively philosophical. They argued about its original condition-was language natural or conventional?-instead of studying its functioning. The categories they established (noun, verb, grammatical gender, etc.) always rest on logical or philosophical grounds. For centuries, from the pre-Socratics to the Stoics and Alexandrians, then during the Aristotelian renaissance, which extended Greek thought until the end of the Latin Middle Ages, language remained an object of speculation, not of observation. Noone was concerned with studying and describing a language for itself, or with inquiring whether the categories of Greek or Latin grammar had general validity. This attitude hardly changed until the eighteenth century. A new phase developed at the beginning of the nineteenth century with the discovery of Sanskrit. It was forthwith discovered that there was a relationship among the languages which have since been called Indo-European. Linguistics was worked out within the framework of comparative grammar with methods which became more and more rigorous as chance findings or decipherings gratified this new science with confirmations of its basic principles and expansions of its range. The work accomplished in a century was considerable and admirable. The methodology tested in the Indo-European field set the pattern. Revised today, it is having new successes. But it should be stated that up until the first decades of our century, linguistics consisted of what was mainly a genetics of languages. It took as its task the study of the evolution of linguistic forms. It set itself up as a historical science, its focus being, everywhere and always, a phase of the history of languages. However, in the midst of these successes, some people were disturbed. What is the nature of a linguistic phenomenon? What is the reality of language? Is it true that it consists of nothing but change? But how does it stay the same and change at the same time? How does it function and what is the relationship of sound to sense? Historical linguistics gave no answers to these questions, having never had to raise them. At the same time, difficulties of a different sort but equally formidable were in the offing. Linguists were beginning to turn to languages which were not written and had no history, especially the Indian languages in America, and they discovered that the traditional frameworks used with the Indo-European languages did not apply to them. They came

A Look at the Development of Linguistics

I9 upon completely different categories which eluded a historical description and called for a new array of definitions and a new method of analysis. Little by little, through many theoretical discussions and under the inspiration of the Cours de linguistique generale of Ferdinand de Saussure (1916), a new notion of language took shape. Linguists came to realize what was incumbent upon them: to study and describe the present linguistic reality by an adequate technique and not to mix any theoretical or historical presupposition with the description, which should be synchronic, and to analyze the language into its own formal elements. Linguistics thus entered into its third phase-the present one. It took for its subject not the philosophy of language nor the evolution of linguistic forms, but brought to the fore the intrinsic reality of language, and aimed at making itself a formal, rigorous, and systematic science. With that, both the historical considerations and the frameworks set up for the Indo-European languages were again called in question. In becoming descriptive, linguistics granted equal importance to all types of languages, written or not, and it had to adapt its methods to them. The problem was indeed to know what a language consisted of and how it functioned. When, following the lead of F. de Saussure, linguists began to envisage language in and of itself, they recognized what was to become the basic principle of modern linguistics: that language forms a system. This holds true for any language, no matter in what culture it is in use, at whatever historical stage we view it. From base to summit, from sounds to the most complex forms of expression, language is a systematic arrangement of parts. It is made up of formal elements put together in variable combinations, according to certain principles of structure. Here is the second key term in linguisticsstructure. By this is meant primarily the structure of the linguistic system gradually revealed, starting from the principle that a language always has a limited number of basic elements, but that these elements, few in number in themselves, yield a large number of combinations. They are, indeed, only grasped within these combinations. Now, methodical analysis leads to the realization that a language actually uses only a small part of the theoretically numerous combinations that would result from these minimal elements being freely assembled. That restriction makes certain specific configurations take shape, which will vary according to the linguistic systems under consideration. Here is what is meant primarily by structure: particular types of relationships articulating the units of a certain level. Each one of the units of a system is thus defined by the relations which it maintains with the other units and by the oppositions into which it enters; as Saussure says, it is a relating and opposing entity. Thus the idea that the data of a language have value in themselves and are objective "facts," absolute

20

PROBLEMS IN GENERAL LINGUISTICS

quantities, susceptible of being considered in isolation, was abandoned. In reality, linguistic entities can be determined only within the system that organizes and governs them, and in terms of each other. They have no value except as elements in a structure. It is first the system which has to be isolated and described. Thus a theory of language as a system of signs and as an arrangement of units in a hierarchy was worked out. It would seem that such an abstract presentation would take us far away from what we call reality. Quite the contrary, it corresponds to the most concrete linguistic experience. The distinctions obtained by analysis agree with those which the speaker practices instinctively. It has been possible to show by experiment that phonemes, that is to say, the distinctive sounds of a language, are psychological realities which the speaker can be led to become conscious of, because when he hears sounds he is in reality identifying phonemes; he recognizes sounds which are sometimes very different as variants of the same phoneme and also recognizes sounds which seem similar as relating to different phonemes. One sees immediately how this approach to linguistics differs from that which formerly prevailed. The positivist notion of the linguistic fact has been replaced by that of relationship. Instead of considering each element by itself and seeking for the fCcause" in an earlier stage, it is envisioned as part of a synchronic totality; "atomism" gives way to "structuralism." By isolating segments of a variable nature and extent within the linguistic data, units of different types are brought out; this leads to characterizing different levels, each one of which is to be described in adequate terms. Hence a great development in the technique and terminology of analysis follows, since all stages have to be explicit. The units of language belong indeed to two different plans; the syntagmatic, when they are taken in a linear relationship within the spoken chain, and the paradigmatic, when they are considered in terms of a possible substitution, each at its level and in its formal class. To describe these relationships, to define these plans, is to refer to the formal structure of the language; and to formalize the description in this way is-without paradox-to make it more and more concrete by reducing the language to the distinctive elements of which it is uniquely constituted and by defining these elements by their mutual dependence. Instead of a series of single, innumerable, and contingent "events," we get a finite number of units and we can characterize a linguistic structure by their distribution and their possible combinations. In proceeding to analyses bearing on different systems, it can be clearly seen that a linguistic form constitutes a definite structure: (I) it is a global unit embracing various parts; (2) these parts are in a formal arrangement that obeys certain constant principles; (3) that which gives the character of a structure

A Look at the Development of Linguistics

21

to the form is that the constituent parts serve a function; (4) finally, these constitutive parts are units on a certain level, in such a way that each unit of a specific level becomes a subunit of the level above. All the essential stages of language are of a discontinuous nature and combine discrete units. It could be said that language is characterized less by what it expresses than by what it distinguishes at all levels : -a distinction of lexemes, which makes it possible to list the notions designated; -a distinction of morphemes, which furnishes the inventory of form classes and subclasses; -a distinction of phonemes, which gives the inventory of nonsignificant phonological distinctions; -a distinction of "merisms" or the features which arrange the phonemes into classes. This is what brings it about that language is a system in which nothing is significant in and of itself, but in which everything is significant as an element of the pattern; structure confers upon the parts their "meaning" or their function. This, too, is what permits unlimited communication; since language is organized systematically and functions according to the rules of a code, the speaker can, with a very small number of basic elements, compose signs, then groups of signs, and finally an unlimited number of utterances, all identifiable for the hearer since the same system exists in him. It can be seen how the notions of system, distinction, and opposition cling closely together and by logical necessity evoke those of dependence and solidarity. There is a solidarity between the terms of an opposition, so that if one of them is affected, the status of the other varies and, as a result, the balance of the system is affected, which can lead to restoring it through the creation of a new opposition at another point. In this respect, each language offers a particular situation at each moment in its history. This consideration has reintroduced the notion of evolution into linguistics today, by specifying diachrony as the relationship between successive systems. Descriptive approach, awareness of system, care taken to extend analysis to the elementary units, and the explicit choice among procedures, such are the features that mark modern linguistics works. To be sure, in practice there are many divergences and conflicts among schools, but we restrict our view to the most general principles, and principles are always more interesting than schools. It has now been discovered that this idea of language had its precursors. It was implicit in the work of the man whom linguists recognize as their first ancestor, the Indian grammarian PaI).ini, who in the middle of the fourth century B.C. had codified the Vedic language into formulae of exemplary

22

PROBLEMS IN GENERAL LINGUISTICS

density; it is a formal, complete, and rigorous description which is not blemished by any speculative or mystical interpretation. But also we ought to do justice to those precursors who were not grammarians and whose work survives, usually anonymous, basic, and ignored, so much a part of our lives at every moment that we do not notice it-I mean the inventors of our modern alphabets. That an alphabet could have been invented and that '''lith a small number of graphic signs everything that is pronounced can be put into writing -this by itself demonstrates at once the articulated structure of language. The Latin and Armenian alphabets are admirable examples of the notation which today is called phonemic. A modern analyst would find hardly anything to change in them; real distinctions are recognized, each letter corresponds to a phoneme, and each phoneme is reproduced by one and the same letter. Alphabetic writing differs in principle from Chinese writing, which is morphemic, or from cuneiform writing, which is syllabic. Those who devised such alphabets for noting down the sounds of their language instinctively recognized-as early phonemicists-that the various sounds that are pronounced come down to a rather limited number of distinctive units which must be represented by as many graphic units. Modern linguists do not operate otherwise when they have to reduce oral languages to writing. We have in these alphabets the most ancient models of analysis: the graphic units of the alphabet and their combination into a great number of specific groups give the closest picture of the structure of the linguistic forms that they reproduce. It is not only linguistic form which comes under this analysis; the function of language must be similarly considered. Language re-produces reality. This is to be understood in the most literal way: reality is produced anew by means of language. The speaker recreates the event and his experience of the event by his discourse. The hearer grasps the discourse first, and through this discourse, the event which is being reproduced. Thus the situation inherent in the practice of language, namely that of exchange and dialogue, confers a double function on the act of discourse; for the speaker it represents reality, for the hearer it recreates that reality. This makes language the very instrument of intersubjective communication. At this point some serious problems immediately arise which we shall leave to the philosophers, notably that of the adequacy of the mind to express "reality." The linguist on his part considers that thought could not exist without language and that as a result the knowledge of the world is being determined by the expression it receives. Language reproduces the world, but by submitting it to its own organization. It is logos, discourse and reason together, as the Greeks saw it. That is what it is by the very fact of being articulated language, consisting in an organic arrangement of parts, in a formal classifica-

A Look at the Development of Linguistics

23

tion of objects and processes. The content to be transmitted (or the "thought") is thus broken down according to a linguistic schema. The "form" of the thought is given its configuration by the structure of the language. And the language in its turn reveals its mediating function in the system of its categories. Each speaker can pose himself as a subject only by implying the other, the partner who, endowed with the same language, shares the same stock of forms, the same syntax for utterance, and the same way of organizing the content. Viewed from the standpoint of the linguistic function, and by virtue of the I-you polarity, individual and society are no longer contradictory but complementary terms. It is, indeed, in and through language that the individual and society define one another. Man has always felt-and poets have often celebrated-the creative power of language, which establishes an imaginary reality, animates inert things, reveals what does not yet exist, and recalls to the present what has disappeared. This is why so many mythologies, having had to explain that at the dawn of time something could be born of nothing, have set up as the creative principle of the world that immaterial and supreme essence, the Word. Indeed there is no higher power, and all the powers of mankind, without exception, spring from it. Society is not possible except through language; nor is the individual. The awakening of consciousness in the child always coincides with the learning of language, which gradually introduces him as an individual into society. But what is the source of this mysterious power that resides in language? Why are the individual and society, together and of the same necessity, grounded in language? Because language represents the highest form of a faculty inherent in the human condition, the faculty of symbolizing. Let us understand by this, very broadly, the faculty of representing the real by a "sign" and of understanding the "sign" as representing the real-the facuIty, then, of establishing a relation of "signification" between one thing and another. Let us consider it first of all in its most general form and apart from language. U sing a symbol is the capacity to identify the characteristic structure of an object and to identify it in various contexts. It is that which is peculiar to man and which makes man a rational creature. The symbolizing faculty, then, permits the formation of the concept as distinct from the concrete object, which is only one realization of the concept. Here is the basis for abstraction as well as the principle of creative imagination. Now, this representative capacity, in essence symbolic, which is at the basis of the conceptual functions, appears only in man. It emerges very early in the child, before language, at the beginning of his conscious life. But it is lacking in animals.

PROBLEMS IN GENERAL LINGUISTICS

Let us, however, make a glorious exception in favour of the bees. According to the memorable observations of K. von Frisch, when a scouting bee has discovered a source of food on her solitary flight, she returns to the hive to announce her find by dancing a special lively dance on the honeycomb, describing certain figures which could be analyzed; she thus indicates to the other bees, who follow slowly behind her, the distance and the direction in which the food is to be found. They then take flight and go unerringly to their objective, which is sometimes quite far from the hive. This is an observation of the highest importance, which seems to indicate that bees communicate among themselves by a particular symbolization and transmit real messages. Should this system of communication be related to the remarkable functioning of the hive? Does the life of social insects imply a certain level of symbolic relationships? It is remarkable even to be able to raise the question. Weare hovering, with hesitation and fascination, on the brink of an immense problem: will man be able for the first time to surmount the biological barrier and see into the interior of an animal society and discover its organizing principle? With this reservation, we can now show more precisely where the difference which separates man from the animals lies. Let us first take great care to distinguish between two notions which are very often confused in speaking of "animal language" -the signal and the symbol. A signal is a physical fact bound to another physical fact by a natural or conventional relationship: lightning heralding a storm, a bell announcing a meal, a cry proclaiming danger. An animal perceives the signal and is capable of reacting adequately to it. It can be trained to identify various signals, that is to say, to connect two sensations through the correlation of the signals. Pavlov's famous conditioned reflexes show this very clearly. Man also, insofar as he is an animal, reacts to signals. But he uses in addition symbols that have been instituted by man; one must learn the meanings of the symbols, one must be able to interpret them in their signifying functions and not simply perceive them as sensory impressions, for symbols have no natural correlation with what they symbolize. Man invents and understands symbols; the animal does not. Everything proceeds from this. The failure to recognize this distinction leads to all sorts of confusions or false problems. It is often said that the trained animal understands human speech. In reality, the animal obeys the spoken word because it has been trained to recognize it as a signal, but it will never be able to interpret it as a symbol. For the same reason, the animal expresses its feelings, but it cannot name them. A beginning or an approximation of language cannot be found in the means of expression used by animals. Between the sensory-motor function and the representative function is a threshold which only human beings have been able to cross. For man was not created twice, once without language and once with lan-

A Look at the Development of Linguistics

guage. The emergence of Homo in the animal series may have been helped by his bodily structure or his nervous organization, but it is due above all to his faculty of symbolic representation, the common source of thought, language, and society. This symbolizing capacity is at the basis of conceptual functions. Thought is nothing other than the power to construct representations of things and to operate on these representations. It is in essence symbolic. 1 The symbolic transformation of the elements of reality or experience into concepts is the method by which the rationalizing power of the mind is brought about. Thought is not a simple reflection of the world; it categorizes reality, and in that organizing function it is so closely associated with language that from this point of view one could be tempted to equate thought and language. Actually the symbolizing faculty in .man attains its supreme realization in language, which is the supreme symbolic expression, all the other systems of communication-graphic, gesticulatory, visual, etc.-being derived from it and presupposing its existence. But language is a special symbolic system organized on two planes. On the one hand it is a physical fact; it makes use of the vocal apparatus for arising and the auditory apparatus for being perceived. In this material aspect it lends itself to observation, description, and recording. On the other hand, it is an immaterial structure, a communication of things signified, which replaces events or experiences by their "evocation." Such is language-a two-sided entity. That is why linguistic symbols are mediatory. They organize thought and they are brought into being in a specific form; they make the internal experience of one person accessible to another in an articulated and representative expression, and not by a signal such as a modulated outcry; they are realized in a given language belonging to a distinct society, not in a vocal utterance common to the whole species. Language offers a model of a relational structure, in the most literal as well as the most comprehensive sense. In discourse it brings words and concepts into contact and in this way, as a substitute for objects and situations, produces signs, which are distinct from their material referents. It establishes those analogical transfers of designations we call metaphors, which are such a powerful factor in the enriching of concepts. It links the propositions in reasoning and becomes the instrument of discursive thought. Finally, language is the most economical symbolism. Unlike other representative systems, it does not demand any muscular effort, it does not involve shifting the body, and it does not require laborious manipulation. Let us imagine what a task it would be to bring the "creation of the world" before the eyes if it were possible to represent it at the expense of an incredible effort by painted, sculpted, or other kinds of figures; then let us see what becomes of the same story when realized in a narrative: just a succession of little vocal

26

PROBLEMS IN GENERAL LINGUISTICS

noises which vanish as soon as they are uttered, as soon as they are heard, but the soul is exalted by them and generations repeat them, and each time that speech discloses the event, the world begins anew. ~o power will ever equal that, and it achieves so much with so little. That such a system of symbols exists reveals to us one of the essential facts -perhaps the most profound-about the human condition; that there is no natural, immediate, and direct relationship between man and the world or between man and man. An intermediary is necessary: this symbolizing apparatus wh\ch has made thought and language possible. Beyond the biological sphere, the capacity for symbolizing is the most characteristic of the human being. I t now remains but to draw the inference from these reflections. In placing man in his relationship with nature or in his relationship with man, by the mediation of language, we establish society. This is not a historical coincidence but an inevitable consequence. For language always realizes itself in a language in a definite and specific linguistic structure which is inseparable from a definite and specific society. Language and society cannot be conceived without each other. Both are given. But both are learned by the human being, who does not possess innate knowledge of them. The child is born and develops in the society of men. It is adult human beings, his parents, who inculcate in him the use of words. The acquisition of language is an experience in the child that goes along with the formation of symbols and the construction of objects. He learns things by their names; he discovers that everything has a name and that learning the names gives him control over the things. But he also discovers that he himself has a name and that by this he can communicate with those around him. Thus there awakens in him the awareness of the social milieu in which he is immersed and which little by little will shape his mind by the intermediary of language. As he becomes capable of more and more complex intellectual operations, he is integrated into the culture which surrounds him. I call culture the human milieu, all that which, beyond the accomplishing of biological functions, gives form, meaning, and content to human life and activity. Culture is inherent in human society, whatever the level of civilization. It consists of a neat number of notions and prescriptions as well as specific prohibitions; what a culture prohibits characterizes it at least as much as what it prescribes. The animal world does not know prohibitions. Now this human phenomenon, culture, is an entirely symbolic phenomenon. Culture is defined as a very complex pattern of representations, organized by a code of relationships and valuestraditions, religion, laws, politics, ethics, arts-everything which man, no matter where he is born, will be steeped in within his deepest consciousness and which will direct his behavior in all forms of his activity; what is this then

A Look at the Development of Linguistics

27

if not a universe of symbols integrated into a specific structure which language reveals and transmits? By means of his language, man assimilates, perpetuates, or transforms his culture. 1\ ow, each culture, as does each language, makes use of a particular set of symbols in \vhich each society is identified. The diversity of languages, the diversity of cultures and their changes, show the conventional nature of the symbolism which links them together. It is definitely the symbol which knots that living cord between man, language, and culture. Such is, in broad outline, the perspective which the recent development in linguistic studies has opened up. Deepening our knowledge of the nature of language, and revealing its relationships with the intelligence as well as with human behavior or the foundations of culture, this investigation is beginning to bring to light the deep-seated functioning of the mind in all its operations. The related sciences are following this progress and cooperate with it on their part by being inspired by the methods and sometimes by the terminology of linguistics. Everything makes for the anticipation that these parallel researches will engender new disciplines and will converge in a real science of culture which will lay the foundations for a theory of the symbolizing activities of man. In addition, it is known that formal descriptions of languages have had a direct usefulness for the construction of logical machines devised for carrying out translations and, conversely, some clarification regarding the way in which thought is coded into language may be expected from the information theory. In the development of this research and these techniques which mark our epoch, we perceive the result of successive symbolizations, more and more abstract, which have their primary and necessary basis in linguistic symbolism. The growing formalization of thought is perhaps leading us to the discovery of an even greater reality. But we could not even conceive of such representations if the structure of language did not contain their first model and, so to speak, their distant harbinger. From Comptes Rendus de l'Academie des Inscriptions et Belles Lettres (1962), Paris, 1963, pp. 369-3 80.

THREE

Saussure After Half a Century

FERDINAND DE SAUSSURE died February 22, 1913. Here we are gathered together fifty years later, on the same date, the twenty-second of February in 1963, for a special commemoration in his city and in his university.! His figure now takes on its authentic character and appears to us in its true grandeur. There is not a linguist today who does not owe him something. There is not a general theory which does not mention his name. Some mystery surrounded his actual life since he so early retired into silence. It is his work we shall deal with. The only fitting eulogy for such a work is that which explains its origins and causes its influence to be understood. Today we see Saussure completely differently from the way his contemporaries could see him. A whole part of him, undoubtedly the most important, was not known until after his death. The science of human speech was gradually changed by him. What is it that Saussure brought to the linguistics of his day, and how has he influenced ours? In order to answer that question, one could go from one of his works to the other, analyzing, comparing, and discussing them. A critical inventory like that is undoubtedly necessary. The admirable and important work of Robert Godel has contributed greatly to this. 2 But this is not our purpose. Leaving to others the detailed description of that work, we shall try to recapture its basic principles in the drive which animated and even formed it. There is in every creative person a certain drive, hidden and permanent, which sustains him and devours him, guides his thoughts, directs him to his task, spurs him on when he lapses, and shows him no quarter when he attempts to evade it. It is not always to be recognized in the different stages, sometimes tentative, in which Saussure's thought was engaged. But once it has been discerned, it illuminates the meaning of his effort and situates him with regard to his predecessors as well as with respect to us. Saussure was first and always a man of fundamentals. He went instinctively to the most basic characteristics which govern the diversity of the empirical datum. In the sphere of language, he suspected the existence of certain

PROBLEMS IN GENERAL LINGUISTICS 30 qualities which were not met with anywhere else. No matter what it was compared to, language always appeared as something different. But how was it different? Considering that activity, human speech, in which so many factors are brought together-biological, physical and psychic, individual and social, historical, aesthetic, and pragmatic-he asked himself, where does language properly belong? A more precise form could be given to this question by connecting it to the two following problems, which we see as being at the center of Saussure's teaching: I. What are the basic data on which linguistics is to be grounded and how can we grasp them? 2. What is the nature of the notions of human speech and by what mode of relationship are they articulated? This preoccupation can be discerned in Saussure from the moment of his entrance into the science, in his Memoire sur ie systeme primitzJ des voyelles dans ies iangues indo-europeens, published when he was twenty-one years old and still one of his titles to fame. The brilliant novice attacked one of the most difficult problems of comparative grammar, a question which, as a matter of fact, did not yet exist and which he was the first to formulate in the appropriate terms. Why did he choose, in a field so vast and promising, such a demanding subject? Let us reread his preface. There he shows that his intention had been to study the Indo-European a but that he had been led to consider "the system of vowels as a whole." This led him to treat "a series of problems of phonetics and morphology, some of which still await their solution and several of which have not even been posed." And as if to excuse himself for having "to cross the most unexplored regions of Indo-European linguistics," he adds this illuminating justification:

If we have nevertheless ventured to go there, convinced in advance that

our inexperience will go astray many times in the labyrinth, it is because, for anyone engaged in these studies, attacking such questions is not temerity, as has often been said, but a necessity. It is the first school through which one must pass, for it is not a matter of transcendental speculations but of the quest for elementary data, without which everything wavers, everything is arbitrariness and incertitude. These last lines could serve as an epigraph to his entire work. They contain the program for his future research, they are portents of his orientation and his goal. Until the end of his life, and more and more insistently-painfully, one might say-the farther he advanced in his thinking, the more he sought for those "elementary data" which constitute language, turning away gradually from the science of his time in which he saw only "arbitrariness and incerti-

Saussure After Half a Century

tude," and this at a period when Indo-European linguistics, confident of its methods, was following the comparative approach with increasing success. It is indeed the elementary data which must be discovered, even (one would like to write, especially) if the purpose is to go back from a historical state of the language to a prehistoric one. Otherwise, the historical development will not have a rational basis, for if there is history, what is it a history of? What changes and what remains the same? How can we say of a linguistic phenomenon considered at two different moments of evolution that it is the same phenomenon? In what does that sameness reside, and, since itlinks for linguists two different objects, how shall we define it? A body of definitions is necessary. We must state the logical connections established within the data, their features, the points of view from which they can be apprehended. To go thus to the fundamentals is the only way-but the sure way-to explain a concrete and contingent fact. In order to reach what is historically concrete, to see the very inevitability of the contingent, we must locate each element in the network of correlations which determines it and admit explicitly that the fact does not exist except by virtue of the definition which we give it. Such was the conviction that forced itself upon Saussure's mind from the beginning and which his whole life was not enough to introduce into linguistic theory. But even if he had formulated then what he was to teach later, he would only have deepened the lack of comprehension or hostility with which his first attempts were met. The chief scholars then, confident of their own truth, refused to listen to this rigorous summons, and the very difficulty of the Memoire was enough to discourage most of them. Perhaps Saussure would have lost heart. It required another generation for his ideas slowly to make their way. It was a fortunate destiny which took him then to Paris. He recovered some confidence in himself, thanks to that remarkable chance which caused him to meet at the same time a sympathetic tutor, Breal, and a group of young linguists like A. Meillet and M. Grammont, who were to be profoundly influenced by his teaching. A new phase of comparative grammar dates from these years in which Saussure inculcated his doctrine, at the same time that he was maturing it, among those who were to develop it. That is why we recall, not only to gauge the personal influence of Saussure but also to assess the progress of the ideas they reveal, the terms of the dedication which Meillet made to his master, Saussure, in 1903, in his Introduction a l'etude comparative des langues indo-europeens, "on the occasion of the twenty-fifth year si~ce the publication of the Memoire ... (r878-1903)." The event would have been more clearly marked if it had been up to Meillet alone; an unpublished letter from Saussure tells us that Meillet had first intended to write "for the anniversary of the publication," from which he was dissuaded by Saussure.

32

PROBLEMS IN GENERAL LINGUISTICS

But even in 1903, that is to say, twenty-five years after the publication of the Memoire of 1878, it still was not possible to know all the prophetic intuitions it contained. Here is a striking example. Saussure had discerned in the Indo-European vowel system several types of a. In point of pure knowledge, the different Indo-European a's are as important as the fundamental particles in nuclear physics. Now one of these a's had the singular quality of behaving differently from the other two. Many discoveries have begun with a similar observation-a disagreement within a system, a disturbance in a field, an abnormal movement in an orbit. Saussure characterized this a by two specific traits. On the one hand, it was related to neither e nor 0; on the other, it was a sonantic coefficient, i.e., it was capable of playing the same double role, vocalic and consonantal, as the nasals and the liquids, and it combined with vowels. Let us note that Saussure spoke of it as a phoneme and not as a sound or an articulation. He did not tell us how this phoneme was pronounced, what sound it approximated in any observable system, or even if it was a vowel or a consonant. The phonic substance was not considered. We are confronted with an algebraic unit, a term of the system, what he would later call a distinguishing and opposing entity. We could not say that this remark attracted much interest even twenty-five years after it had been published. It required another twenty-five years for it to compel recognition, under circumstances which the most audacious imagination could not have conceived. In 1927, the phoneme that had been defined fifty years before by Saussure as the IndoEuropean son antic phoneme was rediscovered by J. Kurylowicz in a historical language, Hittite, which had just then been deciphered, under the guise of the sound written as IJ. This admirable observation made a reality of the theoretical entity postulated by the argument of 1878. Naturally, the phonetic realization of this entity as IJ in Hittite brought a new element into the debate, but it was of a different nature. From then on, two directions in research appeared. Some endeavored to push further the theoretical investigation and to bring to light, especially in Indo-European, the effects and combinations of this "sonantic coefficient." Today one feels that this phoneme is not unique, that it represents a whole class of phonemes, called "laryngeals," which are unevenly distributed among the historical languages. Other linguists emphasize, on the contrary, the descriptive analysis uf these sounds. By this they seek to define their phonetic features, and since the number of laryngeals is still a matter for discussion, from one year to another interpretations can be seen multiplying and giving rise to new controversies. This problem is today at the center of the theory of Indo-European; it is a matter of passionate interest to the diachronists as well as the descriptivists. All this bears witness to the fruitfulness of Saussure's views, which have been fulfilled only in these last decades, half a century after they were

Saussure After Half a Century

33

published. Even those modern linguists who have not read the Memoire are still indebted to it. Here, then, was Saussure, advancing quite young in his career, with the mark of greatness on him. Welcomed with favor at the Ecole des Hautes Etudes, where he immediately found disciples whom his thought enchanted and inspired, and at the Societe de Linguistique where Breal soon made him assistant secretary, a comfortable career opened before him, and everything seemed to herald a long train of discoveries. This expectation was not deceived. Let us recall only the fundamental articles on Baltic intonation, which show the depth of his analysis and remain models for whoever tackles the same research. It is a fact, however, which has been stressed-and deplored-by those who have had to speak of the Saussure of these years, that soon his productivity slowed down. He confined himself to some more and more infrequent articles, and then only to meet the requests of his friends. When he went back to Geneva to take up a chair at the university, he gradually ceased completely to write. Nevertheless, he never ceased to work. What was it that deterred him from publishing? We are beginning to find out. This silence hid a drama which must have been painful; it was aggravated with the years and never had an outcome. It was partly the result of personal circumstances, on which the testimony of his family and friends could shed some light. It was above all a drama of the mind. Saussure alienated himself from his period in the same proportion as he made himself more and more master of his own truth, for that truth made him reject everything that was then taught on the subject of language. But, while he was hesitating to undertake that radical revision which he felt was necessary, he could not bring himself to publish the slightest note if he was not assured first of the fundamental bases of the theory. The depths to which this trouble reached and how many times he was close to becoming discouraged is revealed in a remarkable document, a passage from a letter to Meillet (J anuary 4, 1894) in which he confides, apropos of his studies on Baltic intonation: But I am very disgusted with all that and with the difficulty there is in general to write ten lines concerning the facts of language which have any common sense. I have for a long time been especially concerned with the logical classification of these facts, with the classification of the points of view from which we treat them, and I see more and more both the immensity of the work which is necessary to show the linguist what he is doing, by reducing each operation to its previously specified category; and at the same time the very great vanity of everything which can ultimately be done in linguistics. It is ultimately only the picturesque aspect of a language, that which makes it different from all others in that it belongs to a certain people having certain origins-it is this almost ethnographic aspect which still holds an

PROBLEMS IN GENERAL LINGUISTICS 34 interest for me, and as it happens, I no longer can take any pleasure in devoting myself up to this study without remorse and in enjoying a particular fact which depends on a particular environment. The absolute ineptness of current terminology, the necessity to reform it, and, in order to do that, to show what sort of subject language in general is, come incessantly to spoil my pleasure in history, although I have no dearer wish than not to have to concern myself with language in general. In spite of myself, this will result in a book, in which I shall, without enthusiasm or passion, explain why there is not a single term used in linguistics to which I grant any meaning whatsoever. And I confess that it is not until after that that I shall be able to take up my work at the point at which I left it. That is a perhaps stupid state of mind, which should explain to Duvau why I have, for example, delayed for more than a year over the publication of an article which does not present any material difficulty-and still I have not succeeded in avoiding expressions which are logically odious because a decidedly radical reform would be necessary for this.3

One can see in what sort of argument Saussure was entangled. The more he probed into the nature of human speech, the less he could be satisfied with the accepted ideas. And so he sought a diversion in studies of ethnolinguistic typology, but he was always brought back to his first obsession. Perhaps it was to keep on escaping from it that he threw himself into that ceaseless quest for anagrams .... But we see today what was at stake: Saussure's drama was going to transform linguistics. The obstacles his thinking encountered were to force him to forge the new dimensions which would reorder the facts of language. From this moment, indeed, Saussure saw that to study a particular language led inevitably to the study of language in general. We believe we can reach a linguistic phenomenon as an objective reality. Actually we can grasp it only from a certain point of view, which it is first necessary to define. Let us cease to believe that in language we have to do with a simple object, existing by itself and capable of being grasped in its totality. The first task is to show the linguist "what he is doing," what preliminary operations he performs unconsciously when he approaches linguistic data. Nothing could be more alien to his period than this concern with logic. Linguists were then absorbed in a huge effort of historical investigation, in organizing the comparative materials, and in building up stocks of etymologies. These great undertakings, although quite useful, did not leave room for theoretical concerns. Thus Saussure was alone with his problems. The immensity of the task to be accomplished and the radical character of the necessary reform were enough to make him hesitate and sometimes become discouraged. Nevertheless he did not give it up. He contemplated a book in which he would say these things, in which he would present his views and undertake the complete overhauling of the theory.

Saussure After Half a Century

35

The book was never written, but it survives in rough sketches, in the form of preparatory notes, remarks tossed out rapidly, and drafts; and when he had to give a course in general linguistics in order to fulfill his obligations to the university, he would take up the same themes and bring them to the point at which we know them. In the linguist of 1910 we find again, in effect, the same purpose which guided the novice of 1880: to establish the bases of linguistics. He rejected the categories and notions which he saw in use everywhere because they seemed to him to be foreign to the proper nature of language. What was that nature? He explained his position briefly in certain of his notes, fragments of a reflection which could neither be abandoned nor completely settled: Elsewhere there are things, certain objects, which one is free to consider afterwards from different points of view. In our case there are, primarily, points of view, right or wrong, but simply points of view, with the aid of which, secondarily, one creates things. These creations happen to correspond to realities when the point of departure is right, or not to correspond to them in the opposite case, but in both cases, no thing, no object, is given for a single instant in itself. Not even when the most material fact is dealt with, one which would seem most obviously defined in itself, as would be a series of vocal sounds. 4 Here is our profession of faith regarding linguistic matter: in other fields one can speak of things from such or such point of view, certain that one will find oneself again on firm ground in the object itself. In linguistics, we deny in principle that there are given objects, that there are things which continue to exist when one passes from one order of ideas to another, and that one can, as a result, allow oneself to consider "things" in several orders, as if they were given by themselves. 5

These reflections explain why Saussure considered it so important to show the linguist "what he is doing." He wished to make people understand the error in which linguistics had been engaged from the time it began the study of human speech as a thing, or as a living organism or as a matter to be analyzed by an instrumental technique, or again, as a free and incessant creation of the human imagination. One must get back to the fundamentals and discover that object which is language, to which nothing can be compared. What then was this object which Saussure set up after having made a clean sweep of all accepted notions? Here we touch upon that which is of prime importance in the Saussurian doctrine, upon a principle which assumes a total intuition of language, total both because it contains the whole of his theory and because it embraces the totality of his subject. This principle is that human speech, no matter from what point of view it is studied, is always a double entity, formed of two parts of which the one has no value without the other. There, it seems to me, is the center of the doctrine, the principle from which

PROBLEMS IN GENERAL LINGUISTICS

proceeds all the array of notions and distinctions that constitute the published course. Everything in language is to be defined in double terms; everything bears the imprint and seal of an opposing duality: -the articulatory/acoustical duality; -the duality of sound and sense; -the duality of the individual and society; -the duality of langue and parole; -the duality of the material and the immaterial; -the duality of the "memorial" (paradigmatic) and the syntagmatic; -the duality of sameness and opposition; -the duality of the synchronic and the diachronic, etc. And, once again, none of these terms thus placed in opposition has value by itself or refers to a substantial reality; each of them takes its value from the fact that it is in opposition to the other: The absolutely final law of language is, we dare say, that there is nothing which can ever reside in one term, as a direct consequence of the fact that linguistic symbols are unrelated to what they should designate, so that a is powerless to designate anything without the aid of b, and the same thing is true of b without the aid of a, or that both have no value except through their reciprocal difference, or that neither has any value, even through a certain part of itself (I suppose, "the root," etc.) other than through this same plexus of eternally negative differences. 6 Since human speech does not present a substance in any of its manifestations but only combined or isolated actions of physiological, psychological, or mental forces; and since, nevertheless, all our distinctions, all our terminology, all our ways of speaking are molded according to that involuntary supposition that there is a substance, one cannot refuse, above all, to recognize that the theory of language will have as its most essential task the disentangling of the real nature of our primary distinctions. It is impossible for us to agree that one has the right to construct a theory without performing this essential labor of definition, although this convenient way seems up to now to have satisfied the linguistic pUblic. 7

Certainly a material phenomenon can be taken as the object of a linguistic analysis, for instance, some meaningless segment of an utterance, considered as a simple production of the vocal apparatus, or even an isolated vowel. It is an illusion to believe that here we have hold of a substance; as a matter of fact, it is only by an operation of abstraction and generalization that we can delimit such a topic of study. Saussure insisted upon this, that only the point of view creates this substance. All the aspects of language which we take as given are the result of logical operations which we employ unconsciously. Let us then become conscious of this. Let us open our eyes to the truth that there is not one single aspect of language which is given without the others and which

Saussure After Half a Century

37

one can place above the others as anterior and primordial. Whence this observation: The more one delves into the material proposed for linguistic study, the more one becomes convinced of this truth, which most particularly-it would be useless to conceal it-makes one pause: that the bond established among things is preexistent, in this area, to the things themselves, and serves to determine them. 8

This thesis appears paradoxical and can cause surprise even today. Some linguists reproach Saussure for a propensity to emphasize paradoxes in the functioning of language. But language is actually the most paradoxical thing in the world, and unfortunate are those who do not see this. The further one goes, the more one feels this contrast between oneness as a category of our perception of objects and the pattern of duality which language imposes upon our thought. The more one penetrates into the mechanism of signification, the better one sees that things do not signify by reason of their substantially being so, but by virtue of the formal features which distinguish them from other things of the same class and which it is incumbent upon us to extract. From these views proceeds the doctrine which the disciples of Saussure put into shape and published. Today scrupulous exegetes devote themselves to the necessary task of restoring the lessons of Saussure in their exact terms, with the help of all the materials they can recover. Thanks to their pains we shall have a critical edition of the Cours de linguistique generale which will not only present us with a faithful image of that teaching as it was transmitted in its oral form, but which will allow the Saussurian terminology to be settled with strict accuracy. This doctrine actually informs in one way or another all of the theoretical linguistics of our time. The influence it has had has been enhanced by the effect of the convergence of Saussurian ideas with those of other theorists. Thus in Russia, Baudouin de Courtenay and his disciple Kruszewski proposed at the same time, but independently, a new concept of the phoneme. They distinguished the linguistic function of the phoneme from its articulatory realization. This teaching approached, on a smaller scale, the Saussurian distinction between langue and parole and assigned a differentiating value to the phoneme. This was the first germ of what has developed into a new branch of linguistics, phonology, the theory of the distinctive functions of phonemes and of the structures of their relationships. When they founded it, N. Trubetskoy and R. Jakobson expressly recognized Saussure as well as Baudouin de Courtenay as their precursors. The structuralist trend w~ich emerged in 1928, and which was soon to assume major importance, thus had its origins in Saussure. Although he never

PROBLEMS IN GENERAL LINGUISTICS

used the term "structure" in a doctrinal sense (and it is a term, moreover, which having had to serve as a banner for very different movements has finally lost any precise meaning), the dependence on Saussure is unquestionable among all those who seek the pattern of the general structure of linguistic systems in the relationship of phonemes among themselves. It might be useful to stress this point relative to one of these structuralist schools, the one which is most characteristically national, the American school, insofar as it derives from Bloomfield. It is not widely enough known that Bloomfield wrote a very laudatory review of the Cours de linguistique generale in which he gave credit to Saussure for the distinction between langue and parole and concluded: "He has given us the theoretical basis for a science of human speech."9 Different as American linguistics has become, it still keeps a link with Saussure. Like all productive thought, the Saussurian concept of language had consequences that were not immediately perceived. There is even a part of his teaching which remained unproductive and inactive for a long time. It is the part which relates to language as a system of signs, and the analysis of the sign into the signifier and the signified. That was a new principle, the one of the two-sided unit. In recent years, linguists have debated the notion of sign: to what extent do the two aspects correspond, how does the unit hold or split through diachrony, etc. Many points of the theory are still to be discussed. Among others, it must be asked if the notion of the sign could be retained as a principle of analysis at all levels. We have stated elsewhere that the sentence as such does not admit of segmentation into units of the type of the sign. But what we wish to emphasize here is the scope of the principle of the sign when set up as a unit of language. From that it results that language becomes a semiotic system: "The task of the linguist," said Saussure, "is to find out what makes language a special system within the mass of semiological data .... But to me the language problem is mainly semiological. "10 We see that this principle is now gaining ground outside linguistics and penetrating into the sciences of man, which are becoming aware of their own semiotics. Far from language being swallowed up in society, it is society which is beginning to recognize itself as "a language." Analysts of society ask themselves if certain social structures, or, in another context, those complex discourses which myths are, should not be considered as signifiers for which one has to search out the signified. These innovating investigations suggest that the basic characteristic of language, that it is composed of signs, could be common to all those social phenomena which constitute culture. It seems to us that one should draw a fundamental distinction between two orders of phenomena: on the one side the physiological and biological data, which present a "simple" nature (no matter what their complexity may be)

Saussure After Half a Century

39 because they hold entirely within the field in which they appear and because their structures form and diversify themselves on successive levels in the order of the same relationships; on the other side, the phenomena belonging to the interhuman milieu, which have the characteristic that they can never be taken as simple data or defined in the order of their own nature but must always be understood as double from the fact that they are connected to something else, whatever their "referent" may be. A fact of culture is such only insofar as it refers to something else. The day when a science of culture takes shape, it will probably be founded upon that chief feature, and it will develop its own dualities on the model Saussure gave for language, without necessarily conforming to it. No science of man will be spared this reflection on its subject and its place within a general science of culture, for man is not born in nature but in culture. What a strange destiny ideas have, sometimes seeming to lead lives of their own, revealing or contradicting or recreating the figure of their creator. One could pause a long time over this contrast: the actual life of Saussure compared with the fortunes of his ideas. A man alone in his thought for almost his whole life, unable to bring himself to teach what he deemed wrong or fanciful, feeling that it was necessary to recast everything but less and less tempted to do it, and finally, after many diversions that could not rescue him from the torment of his personal truth, imparting to a few students some ideas on the nature of language which never seemed to him to be matured enough to be published. He died in 1913, little known outside the narrow circle of his pupils and a few friends, even then almost forgotten by his contemporaries. Meillet, in the moving obituary he then devoted to him, deplored the fact that Saussure's life ended on an unfinished work: "After more than thirty years, the ideas expressed by Ferdinand de Saussure in his early work have not exhausted their vitality. And nevertheless his disciples have the feeling that he has not held anything like the place in the linguistics of his time which his brilliant gifts deserved." 11 And he ended with this poignant regret: "[Saussure] had produced the finest book of comparative grammar that has ever been written, he had sown ideas and laid down firm theories; he had left his mark on numerous pupils, and still he had not completely fulfilled his destiny."12 Three years after the death of Saussure, the Cours de linguistique generale appeared, edited from the notes of students by Bally and Sechehaye. In 1916, amidst the clash of arms, who could have been concerned with a book on linguistics? Nietzsche's saying, that great events arrive on dove's feet, was never truer. Today, fifty years have passed since the death of Saussure, two generations separate us from him, and what do wesee? Linguistics now holds an important

PROBLEMS IN GENERAL LINGUISTICS

place among the sciences dealing with man and society and is among the most active of them in theoretical research as well as in technical developments. And this new linguistics has its origins in Saussure, and it is in Saussure that it finds its bearings and takes shape. Among all the currents that cross it, among all the schools into which it is divided, the innovating role of Saussure is proclaimed. That seed of brightness, culled by a few disciples, has become a great light which reveals a landscape filled with his presence. What we are here asserting is that Saussure belongs henceforth to the history of European thought. A forerunner in doctrines which in the past fifty years have transformed the theory of language, he has opened up unforgettable vistas on the highest and most mysterious faculty of man. At the same time, in placing on the horizon of science and philosophy the notion of "sign" as a bilateral unit, he has contributed to the advent of formal thought in the sciences of society and culture, and to the founding of a general semiology. Taking in at a glance the half-century that has just elapsed, we can say that Saussure really did accomplish his destiny. Beyond his earthly life, his ideas spread further than he could have imagined, and his posthumous destiny has become, so to speak, a second life which henceforth mingles with ours.

From Cahiers Ferdinand de Saussure

20

(1963) : 7-21

COIllIllunication

FOUR

The Nature of the Linguistic Sign

THE IDEA OF THE linguistic sign, which is today asserted or implied in most works of general linguistics, came from Ferdinand de Saussure. And it was as an obvious truth, not yet explicit but nevertheless undeniable in fact, that Saussure taught that the nature of the sign is arbitrary. The formula immediately commanded attention. Every utterance concerning the essence of language or the modalities of discourse begins with a statement of the arbitrary character of the linguistic sign. The principle is of such significance that any thinking bearing upon any part of linguistics whatsoever necessarily encounters it. That it is cited everywhere and always granted as obvious are two good reasons for seeking at least to understand the sense in which Saussure took it and the nature of the proofs which show it. In the Cours de linguistique generale, 1 this definition is explained in very simple statements. One calls sign "the total resultant of the association of a signifier [=sound image] and what is signified [=concept] ... " "The idea of 'sister' is not linked by any inner relationship to the succession of sounds s-o-r which serves as its signifier in French; that it could be represented equally by just any other sequence is proved by differences among languages and by the very existence of different languages: the signified 'ox' has as its signifier b-o-f on one side of the border and o-k-s (Ochs) on the other" (p. 102 [pp. 67-68]). This ought to establish that "The bond between the signifier and the signified is arbitrary," or, more simply, that "the linguistic sign is arbitrary" [po 67]. By "arbitrary," the author means that "it is unmotivated, i.e., arbitrary in that it actually has no natural connection with the signified" (p. 103 [po 69]). This characteristic ought then to explain the very fact by which it is verified: namely, that expressions of a given notion vary in time and space and in consequence have no necessary relationship with it. We do not contemplate discussing this conclusion in the name of other principles or by starting with different definitions. The question is whether it is consistent and whether, having accepted the bipartite nature of the sign (and we do accept it), it follows that the sign should be characterized as

PROBLEMS IN GENERAL LINGUISTICS 44 arbitrary. It has just been seen that Saussure took the linguistic sign to be made up of a signifier and signified. Now-and this is essential-he meant by "signifier," the concept. He declared in so many words (p. 100 [po 66]) that the "linguistic sign unites, not a thing and a name, but a concept and a sound image." But immediately afterward he stated that the nature of the sign is arbitrary because it "actually has no natural connection with the signified" [po 69]. It is clear that the argument is falsified by an unconscious and surreptitious recourse to a third term which was not included in the initial definition. This third term is the thing itself, the reality. Even though Saussure said that the idea of "sister" is not connected to the signifier s-o-r, he was not thinking any the less of the reality of the notion. When he spoke of the difference between b-o-f and o-k-s, he was referring in spite of himself to the fact that these two terms applied to the same reality. Here, then, is the thing, expressly excluded at first from the definition of the sign, now creeping into it by a detour, and permanently installing a contradiction there. For if one states in principle-and with reason-that language is form, not substance (p. 163 [po II3]), it is necessary to admit-and Saussure asserted it plainlythat linguistics is exclusively a science of forms. Even more imperative is the necessity for leaving the "substance," sister or ox, outside the realm of the sign. Now it is only if one thinks of the animal ox in its concrete and "substantial" particularity, that one is justified in considering "arbitrary" the relationship between bof on the one hand and oks on the other to the same reality. There is thus a contradiction between the way in which Saussure defined the linguistic sign and the fundamental nature which he attributed to it. Such an anomaly in Saussure's close reasoning does not seem to me to be imputable to a relaxation of his critical attention. I would see instead a distinctive trait of the historical and relativist thought of the end of the nineteenth century, an inclination often met with in the philosophical reflection of comparative thought. Different people react differently to the same phenomenon. The infinite diversity of attitudes and judgments leads to the consideration that apparently nothing is necessary. From the universal dissimilarity, a universal contingency is inferred. The Saussurian concept is in some measure dependent on this system of thought. To decide that the linguistic sign is arbitrary because the same animal is called bceuf in one country and Ochs elsewhere, is equivalent to saying that the notion of mourning is arbitrary because in Europe it is symbolized by black, in China by white. Arbitrary, yes, but only under the impassive regard of Sirius or for the person who limits himself to observing from the outside the bond established between an objective reality and human behavior and condemns himself thus to seeing nothing in it but contingency. Certainly with respect to a same

The Nature of the Linguistic Sign

45

reality, all the denominations have equal value; that they exist is thus the proof that none of them can claim that the denomination in itself is absolute. This is true. It is only too true and thus not very instructive. The real problem is far more profound. It consists in discerning the inner structure of the phenomenon of which only the outward appearance is perceived, and in describing its relationship with the ensemble of manifestations on which it depends. And so it is for the linguistic sign. One of the components of the sign, the sound image, makes up the signifier; the other, the concept, is the signified. Between the signifier and the signified, the connection is not arbitrary; on the contrary, it is necessary. The concept (the "signified") ba:uf is perforce identical in my consciousness with the sound sequence (the "signifier") boj. How could it be otherwise? Together the two are imprinted on my mind, together they evoke each other under any circumstance. There is such a close symbiosis between them that the concept of bceuf is like the soul of the sound image boJ. The mind does not contain empty forms, concepts without names. Saussure himself said: Psychologically our thought-apart from its expression in words-is only a shapeless and indistinct mass. Philosophers and linguists have always agreed in recognizing that without the help of signs we would be unable to make a clear-cut, consistent distinction between two ideas. Without language, thought is a vague, uncharted nebula. There are no preexisting ideas, and nothing is distinct before the appearance of language (p. 161 [pp. I I I 112]).

Conversely, the mind accepts only a sound form that incorporates a representation identifiable for it; if it does not, it rejects it as unknown or foreign. The signifier and the signified, the mental representation and the sound image, are thus in reality the two aspects of a single notion and together make up the ensemble as the embodier and the embodiment. The signifier is the phonic translation of a concept; the signified is the mental counterpart of the signifier. This consubstantiality of the signifier and the signified assures the structural unity of the linguistic sign. Here again we appeal to Saussure himself for what he said of language: Language can also be compared with a sheet of paper: thought is the front and the sound the back; one cannot cut the front without cutting the back at the same time; likewise in language, one can neither divide sound from thought nor thought from sound; the division could be accomplished only abstractedly, and the result would be either pure psychology or pure phonology (p. 163 [po 113]).

PROBLEMS IN GENERAL LINGUISTICS

What Saussure says here about language holds above all for the linguistic sign in which the primary characteristics of language are incontestably fixed. One now sees the zone of the "arbitrary," and one can set limits to it. What is arbitrary is that one certain sign and no other is applied to a certain element of reality, and not to any other. In this sense, and only in this sense, is it permissible to speak of contingency, and even in so doing we would seek less to solve the problem than to point it out and then to take leave of it temporarily. For the problem is none other than the famous CPVOet or (Jioet? and can only be resolved by decree. It is indeed the metaphysical problem of the agreement between the mind and the world transposed into linguistic terms, a problem which the linguist will perhaps one day be able to attack with results but which he will do better to put aside for the moment. To establish the relationship as arbitrary is for the linguist a way of defending himself against this question and also against the solution which the speaker brings instinctively to it. For the speaker there is a complete equivalence between language and reality. The sign overlies and commands reality; even better, it is that reality (nomen/omen, speech taboos, the magic power of the word, etc.). As a matter of fact, the point of view of the speaker and of the linguist are so different in this regard that the assertion of the linguist as to the arbitrariness of designations does not refute the contrary feeling of the speaker. But, whatever the case may be, the nature of the linguistic sign is not at all involved if one defines it as Saussure did, since the essence of this definition is precisely to consider only the relationship of the signifier and the signified. The domain of the arbitrary is thus left outside the extension of the linguistic sign. It is thus rather pointless to defend the principle of the "arbitrariness of the sign" against the objection which could be raised from onomatopoeia and expressive words (Saussure, pp. 103-104 [pp. 69-70]). Not only because their range of use is relatively limited and because expressivity is an essentially transitory, subjective, and often secondary effect, but especially because, here again, whatever the reality is that is depicted by the onomatopoeia or the expressive word, the allusion to that reality in most cases is not immediate and is only admitted by a symbolic convention analogous to the convention that sanctions the ordinary signs of the system. We thus get back to the definition and the characteristics which are valid for all signs. The arbitrary does not exist here either, except with respect to the phenomenon or to the material object, and does not interfere with the actual composition of the sign. Some of the conclusions which Saussure drew from the principle here discussed and which had wide effect should now be briefly considered. For instance, he demonstrated admirably that one can speak at the same time of the mutability and immutability of the sign; mutability, because since it is

The Nature of the Linguistic Sign

47

arbitrary' it is always open to change, and immutability, because being arbitrary it cannot be challenged in the name of a rational norm. "Language is radically powerless to defend itself against the forces which from one moment to the next are shifting the relationship between the signified and the signifier. This is one of the consequences of the arbitrary nature of the sign" (p. 112 [po 75]). The merit of this analysis is in no way diminished, but on the contrary is reinforced, if one states more precisely the relationship to which it in fact applies. It is not between the signifier and the signified that the relationship is modified and at the same time remains immutable; it is between the sign and the object; that is, in other terms, the objective motivation of the designation, submitted, as such, to the action of various historical factors. What Saussure demonstrated remains true, but true of the signification, not the sign. Another problem, no less important, which the definition of the sign concerns directly, is that of value, in which Saussure thought to find a confirmation of his views: Cc. • • the choice of a given slice of sound to name a given idea is completely arbitrary. If this were not true, the notion of value would be compromised, for it would include an externally imposed element. But actually values remain entirely relative, and that is why the bond between the sound and the idea is radically arbitrary" (p. 163 [po 113]). It is worth the trouble to take up in succession the several parts of this argument. The choice that invokes a certain sound slice for a certain idea is not at all arbitrary; this sound slice would not exist without the corresponding idea and vice versa. In reality, Saussure was always thinking of the representation of the real object (although he spoke of the CCidea") and of the evidently unnecessary and unmotivated character of the bond which united the sign to the thing signified. The proof of this confusion lies in the following sentence in which I have underlined the characteristic part: cclf this were not true, the notion of value would be compromised since it would include an externally imposed element." It is indeed an CCexternally imposed element," that is, the objective reality which this argument takes as a pole of reference. But if one considers the sign in itself and insofar as it is the carrier of value, the arbitrary is necessarily eliminated. For-the last proposition is the one which most clearly includes its own refutation-it is quite true that values remain entirely ccrelative" but the question is how and with respect to what. Let us state this at once: value is an element of the sign; if the sign taken in itself is not arbitrary, as we think to have shown, it follows that the ccrelative" character of the value cannot depend on the "arbitrary" nature of the sign. Since it is necessary to leave out of account the conformity of the sign to reality, all the more should one consider the value as an attribute only of the form, not of the substance. From then on, to say that the values are C!relative" means that

PROBLEMS IN GENERAL LINGUISTICS

they are relative to each other. Now, is that not precisely the proof of their necessity? We deal no longer here with the isolated sign but with language as a system of signs, and no one has conceived of and described the systematic economy of language as forcefully as Saussure. Whoever says system says arrangement or conformity of parts in a structure which transcends and explains its elements. Everything is so necessary in it that modifications of the whole and of details reciprocally condition one another. The relativity of values is the best proof that they depend closely upon one another in the synchrony of a system which is always being threatened, always being restored. The point is that all values are values of opposition and are defined only by their difference. Opposed to each other, they maintain themselves in a mutual relationship of necessity. An opposition is, owing to the force of circumstances, subtended by necessity, as it is necessity which gives shape to the opposition. If language is something other than a fortuitous conglomeration of erratic notions and sounds uttered at random, it is because necessity is inherent in its structure as in all structure. It emerges, then, that the role of contingency inherent in language affects denomination insofar as denomination is a phonic symbol of reality and affects it in its relationship with reality. But the sign, the primordial element in the linguistic system, includes a signifier and a signified whose bond has to be recognized as necessary, these two components being consubstantially the same. The absolute character of the linguistic sign thus understood commands in its turn the dialectical necessity of values of constant opposition, and forms the structural principle of language. It is perhaps the best evidence of the fruitfulness of a doctrine that it can engender a contradiction which promotes it. In restoring the true nature of the sign in the internal conditioning of the system, we go beyond Saussure himself to affirm the rigor of Saussure's thought.

From Acta Linquistica 1 (Copenhagen, 1939) : 23-29

FIVE

Animal Communication and Human Language

To APPLY THE NOTION OF language to the animal world is admissible only at the price of misusing terms. We know that it has been impossible until now to prove that animals enjoy, even in a rudimentary form, a means of expression endowed with the characteristics and functions of human speech. All serious observations made of animal communities, all attempts to establish or verify, by means of various technical devices, any form of speech comparable to that of man have failed. It does not seem that animals which emit certain kinds of calls are thereby displaying any behavior from which we may infer that they are conveying "spoken" messages to one another. The fundamental conditions for a strictly linguistic communication seem to be lacking even in the higher animal world. The case of the bees, however, is different. At any rate, it has become apparent lately that it may turn out to be different. Everything confirms the belief that the bees possess the means of communicating with one another-a fact which has been observed for a long time. The amazing organization of their colonies, the differentiation and coordination of their activities, their capacity for reacting collectively to unforeseen circumstances, lead us to suppose that they are capable of exchanging real messages. The attention of observers has been drawn particularly to the way in which the bees are informed when one of them has discovered a source of food. Consider, e.g., a foraging bee discovering on its flight a sugar solution, placed at a certain point experimentally in order to attract its attention. It will drink of it, and while it feeds, the experimenter carefully puts a mark on it. Then it flies back to the hive. A few seconds later a flight of bees arrives on the spot, all from the same hive. The bee which discovered the food is not among them. It must have informed the others, and the information must have been quite precise since they are able to reach the spot without any guide, although it often is at a considerable distance from the hive and always out of the bees' sight. There is no error or hesitation in locating it. If the foraging bee has chosen one particular flower among others which could have also attracted it, then the

50

PROBLEMS IN GENERAL LINGUISTICS

bees arriving on the scene after its return fly to the same flower, neglecting all others. It seems clear that the scouting bee has indicated to its fellow bees the spot whence it has come. But how? This fascinating problem has baffled observers for a long time. We owe it to Karl von Frisch (professor of zoology at the University of Munich) and to the experiments he conducted for some thirty years, that we are now in the possession of principles which enable us to solve the problem. His research has revealed the method of communication among bees. Working with a transparent hive, he has observed the conduct of the bee returning after the discovery of honey. It is immediately surrounded by the others. The excitement in the hive is great. They stretch out their antennae toward it to collect the pollen with which it is laden or they drink the nectar which it disgorges. Then, followed by the others, the scouting bee proceeds to perform dances. This is the critical moment and constitutes the act of communication. The bee performs two different dances, according to the kind of information it intends to convey. In the one dance it traces horizontal circles from right to left, then from left to right, in succession (round dance). In the other dance (wagging dance) it wags its abdomen continually and cuts what appears to be a figure eight in the following manner: it flies straight, then makes a full left turn, flies straight again, and begins a full turn to the right, etc. After the dances, one or several bees leave the hive and go straight to the supply spot visited by the first bee. Once they have had their fill they regain the hive, where they, in turn, perform the same dances. This causes fresh departures so that, after a few coming and goings, some hundreds of bees swarm to the spot where the forager discovered the food. The round dance and the wagging dance, then, appear to be the actual message which announces the discovery to the hive. The difference between the two dances still awaited an explanation. Frisch thought that it refers to the nature of the food: the round dance announcing the nectar, the wagging dance the pollen. These facts and their interpretation, first presented in 1923, have been much publicized, and even popularized, in the meantime.! It is easy to appreciate the lively interest which they have aroused. Nevertheless, they do not entitle us to ascribe to the bees a language in the strict sense of the word. This position, however, was changed completely as a result of further experiments by Karl von Frisch, extending and correcting his first observations. He announced his findings in 1948 in technical journals and summarized them in 1950 in a small volume presenting a series of lectures he had delivered in the United States. 2 After conducting, literally, thousands of experiments with truly admirable patience and ingenuity, he succeeded in determining the real meaning of the dances. The essential new information

Animal Communication and Human Language

51

which he provided is that the dances indicate the distance from the hive to the food and not, as he thought at first, the nature of the food. The round dance announces that the food site must be sought close by within the radius of approximately a hundred meters from the hive. The bees fly out hovering not far from the hive until they have found the spot. The other dance performed by the foraging bee, in which it wags its abdomen and cuts figures of eight, indicates that the point is at a greater distance, between a hundred meters and six kilometers. This message contains two distinct pieces of information, one about the distance, the other about the direction. The distance is indicated by the number of figures traced in a given time. It varies always in inverse proportion to their frequency. For example, the bee describes nine to ten complete cycles of the dance in fifteen seconds when the distance is a hundred meters, seven for two hundred meters, four and a half for one kilometer, and only two for six kilometers. The greater the distance, the slower the dance. As for the direction in which the food is to be sought, this is indicated by the axis of the figure eight and its relation to the sun. According to its inclination right or left this axis gives the angle which the site forms with the sun. By virtue of their particular sensitiveness to polarized light, the bees are capable of finding their bearings even when the sky is overcast. In practice there exist slight variations, in measuring the distance, between one bee and another or between one hive and another, but the variations do not affect the choice of the dance. This interpretation is the result of approximately four thousand experiments which other zoologists, at first inclined to be sceptical, have repeated and confirmed in Europe and in the United States. 3 We now have the means of ascertaining that it is in fact the dance with its two variations which the bees use to inform their fellow bees about a discovery and to guide them to the spot by giving information about direction and distance. The nature of the food, furthermore, is disclosed to the other bees by the scent on the scouting bee or by the nectar which it has drunk and which they now absorb from it. Then they take wing and infallibly reach the spot. The experimenter thus can predict the behavior of the hive and verify the information given, according to the type and rhythm of the dance. The importance of these discoveries for the study of animal psychology need not be stressed. We should like to dwell here on a less obvious aspect of the problem, which Frisch, intent on describing objectively his experiments, has not touched on. We are, for the first time, in a position to ascertain with precision the methods of communication used in an insect colony. We can, likewise, for the first time envisage the working of an animal "language." It may be well to examine briefly if and in what sense it can or cannot be called a language and how these observations on the bees could help us to find, by contrast or resemblance, a definition of human speech.

52

PROBLEMS IN GENERAL LINGUISTICS

The bees appear to be capable of giving and receiving real messages which contain several data. They can register reports concerning the position and distance of a certain object. They can store these data in some kind of "memory." They can, furthermore, communicate them by means of symbols, using different somatic movements. Indeed, the most remarkable thing is that they show an aptitude for symbolizing: there is undoubtedly a "conventional" relation between their behavior and the facts it conveys. This relation is perceived by the other bees in the terms in which it is transmitted to them and becomes an actuating force. So far we find among bees the very conditions without which no language is possible, i.e., the capacity for formulating and interpreting a "sign" which refers to a certain "reality," the memory of an experience undergone, and the ability to decompose the remembered experience. The message passed on contains three items of information, or, more precisely, only three have been identified until now: the existence of a source of food, its distance, and its direction. These elements could be arranged in a somewhat different way. The round dance indicates simply the presence of food and merely implies the fact that it is not far away. It is based on the mechanical principle of "all or nothing." The other dance conveys a real communication. The existence of food, this time, is implicit in two data (distance and direction) which are explicitly announced. There are thus several points of resemblance with human language. An effective, though rudimentary, symbolism is brought into play. Through it objective data are turned into formalized gestures conveying variable elements and an invariable "meaning." We are faced here with a language in the strict sense of the term, considering not only the way it functions but also the medium in which it takes place: the system is operative within a given community, and each member of the community is capable of using and of understanding it. But the differences between the bee language and human language are considerable, and they help us to understand the truly distinctive characteristics of the latter. It should be noted, above all, that the bee's message consists entirely of physical motion, a dance, without the intervention of any "vocal" organ, whereas there can be no real language without the exercise of voice. This leads us to another difference of a physical nature. Effectuated as it is without the exercise of voice, by means of gestures only, communication between bees necessarily occurs under conditions which permit visual perception, i.e., in daylight. It cannot be made effective in darkness. Human language is not subject to this limitation. A very important difference exists, furthermore, with regard to the circumstances in which the communication is made. The bee's message does not call for any reply from those to whom it is addressed, except that it evokes a

Animal Communication and Human Language

53

particular behavior which is not strictly an answer. This means that the language of the bees lacks the dialogue which is distinctive of human speech. \Ve speak to others who speak to us: such is the nature of human intercourse. This reveals yet another contrast. Because the bees are incapable of dialogue, the communication concerns only a certain objective fact. No "linguistic" information is involved, there being no reply. For a reply is a linguistic reaction to a linguistic manifestation. l\loreover, the bee's message cannot be reproduced by another bee which has not seen for itself what the first bee has announced. There is no indication, for example, that a bee goes off to another hive with the message it has received in its own hive. This would constitute a kind of transmission or relay. Human language is different; for in the dialogue the reference to the objective experience and the reaction to its linguistic manifestation mix freely and without limitation. The bee does not construe a message from another message. Each bee, once advised by the scouting bee's dance, flies out and feeds at the spot indicated, reproducing the same information on its return, not with reference to the first message but with reference to the fact it has just verified itself. Now the characteristic of language is to produce a substitute for experience which can be passed on ad infinitum in time and space. This is the nature of our symbolism and the basis of linguistic tradition. If we now consider the content of the message it is easy to see that it always concerns only one fact, namely, food, and that the only variations of this theme concern the question of space. The contrast with the boundless possibilities of human language is obvious. Furthermore, the behavior which expresses the bee's message is a special form of symbolism. It consists in tracing off an objective situation of fact, the only situation which can be translated into a message, without any possibility of variation or transposition. In human language, on the contrary, the symbol as such does not trace out the facts of experience in the sense that there is no necessary relationship between the objective reference and the linguistic form. Many more distinctions could be made here from the standpoint of human symbolism, the nature and function of which have as yet been little studied. But the difference is already sufficiently indicated. Finally, one more feature of the communication among bees should be mentioned which distinguishes it sharply from human language. The bee's message cannot be analyzed. We can see in it only an overall reference to a total content; the only possible differentiation pertains to the spatial position of the reported object. But it is impossible to resolve this content into its constituent "morphemes" and to make each morpheme correspond to an element of what has been enounced. This is precisely where the distinctive character of human speech manifests itself. Each enunciation made by man

54

PROBLEMS IN GENERAL LINGUISTICS

can be reduced to elements which combine easily and freely according to definite laws so that a small number of morphemes admits of a great number of combinations. Hence proceeds the variety of human language-which has the capacity of expressing everything. A more searching analysis of language reveals that this restricted number of morphemes, or elements of meaning, can be reduced to even less numerous "phonemes," or elements of articulation, devoid of meaning. It is the selective and distinctive grouping of these elements of articulation which produces the sense units. These "empty" phonemes, organized in systems, constitute the basis of every language. It is evident that no such constituent parts can be isolated in the language of the bees. It cannot be reduced to identifiable and distinctive elements.' All these observations bring out the essential difference between the method of communication discovered among bees and our human language. This difference can be stated summarily in one phrase which seems to give the most appropriate definition of the manner of communication used by the bees: it is not a language but a signal code. All the characteristics of a code are present: the fixity of the subject matter, the invariability of the message, the relation to a single set of circumstances, the impossibility of separating the components of the message, and its unilateral transmission. Nevertheless, it is significant that this code, the only form of language found so far among animals, is the property of insects which live in a society. Society is likewise the condition of human language. One of the most interesting aspects of the discoveries of Karl von Frisch is that, apart from the insights into the life of the insect world, he has indirectly enlightened us as to the conditions of human language and its underlying symbolism. It is likely that further progress of this research will bring a further penetration of the possibilities and nuances of this form of communication. But the mere discovery of its existence, its nature, and its way of functioning is a contribution toward a better understanding of the origins of language and the definition of man. 5

From Diogenes, no. 1 (1952), pp. 1-7

SIX

Categories of Thought and Language

WE USE THE LANGUAGE we speak in infinitely varied ways, a simple enumeration of which would be coextensive with a list of the activities in which the human mind can engage. In their diversity, these uses have, however, two characteristics in common. One is that the reality of language, as a general rule, remains unconscious; except when language is especially studied for itself, we have no more than a very faint and fleeting awareness of the operations which we accomplish in order to talk. The other is that, no matter how abstract or how specialized the operations of thought may be, they receive expression in language. We can say everything, and we can say it as we wish. From this proceeds the conviction, widely prevalent and itself unconscious, as is everything that regards human speech, that thinking and speaking are activities distinct by nature, associated for the practical necessity of communication, but which both have their respective domain and their independent possibilities, those of language consisting of the resources offered to the mind for what is called the expression of thought. Such is the problem which we are considering briefly here, for the special purpose of clearing up some ambiguities for which the very nature of human speech is responsible. Certainly speech, being spoken, is used to convey "what we want to say." But CCwhat we want to say" or "what we have in mind" or "our thought" or whatever name it is designated by is the content of thought, very difficult to define in itself, except by the characteristics of intention or as a psychic structure, etc. This content receives form when it is uttered, and only thus. It receives form from language and in language which is the mold for all possible expression; it cannot be dissociated from it and it cannot transcend it. Now this language has a configuration in all its parts and as a totality. It is in addition organized as an arrangement of distinct and distinguishing CCsigns," capable themselves of being broken down into inferior units or of being grouped into complex units. This great structure, which includes substructures of several levels, gives its form to the content of thought. To become transmissible, this content must be distributed among morphemes of certain classes, arranged in

PROBLEMS IN GENERAL LINGUISTICS

a certain order, etc. In short, this content has to pass through language and conform to its framework. Otherwise thought amounts, if not exactly to nothing, at least to something so vague and so undifferentiated that we have no means for comprehending it as "content" distinct from the form conferred upon it by language. Linguistic form is not only the condition for transmissibility, but first of all the condition for the realization of thought. We do not grasp thought unless it has already been adapted to the framework of language. Without that, there is only obscure volition, impulse venting itself in gestures, or mimicry. That is to say that the question of whether thought can do without language or skirt it like an obstacle emerges as meaningless as soon as one analyzes with any rigor the terms of the problem. This is, however, still only a de/acto relationship. To set up these two terms, thought and language, as interdependent and mutually necessary does not say how they are interdependent and why they are judged to be indispensable to each other. Between a thought that can only be materialized in language and a language that has no other function than to "signify," one would wish to state a specific relationship, for it is obvious that the terms in question are not symmetrical. To speak of the container and the contents is to simplify. The image should not delude us. Strictly speaking, thought is not matter to which language lends form, since at no time could this "container" be imagined as empty of its contents, nor the "contents" as independent of their "container." And so the question becomes the following: while granting absolutely that thought cannot be grasped except as formed and made a reality in language, have we any means to recognize in thought such characteristics as would belong to it alone and owe nothing to linguistic expression? We can describe language by itself. It would be necessary in the same way to apprehend thought directly. If it were possible to define thought by features belonging to it exclusively, it would be seen at once how it accommodates itself to language and what the nature of their relationship is. It might be convenient to approach the problem by way of "categories," which appear as intermediaries. They present different aspects, depending on whether they are categories of thought or language. This difference might shed light on their respective natures. For example, we immediately perceive that thought can freely specify its categories and invent new ones, while linguistic categories, as attributes of a system which each speaker receives and maintains, are not modifiable according to each person's whim. We also see this other difference: that thought can claim to set up universal categories but that linguistic categories are always categories of a particular language. At first sight, this would confirm the preeminent and independent position of thought with regard to language. We cannot, however, as so many authors have done, simply pose the ques·

Categories of Thought and Language

57 tion in such general terms. We must enter into a concrete historical situation, and study the categories of a specific thought and a specific language. Only on this condition will we avoid arbitrary stands and speculative solutions. Now, we are fortunate to have at our disposal data which one would say were ready for our examination, already worked out and stated objectively within a well-known system: the Aristotle's categories. In the examination of these categories, we may dispense with philosophical technicalities. We will consider them simply as an inventory of properties which a Greek thinker thought could be predicated of a subject and, consequently, as the list of a priori concepts which, according to him, organize experience. It is a document of great value for our purpose. Let us recall at first the fundamental text, which gives the most complete list of these properties, ten in all (Categories 4 1): Each expression when it is not part of a combination means: the substance, or how much, or of what kind, or relating to what, or where, or when, or to be in a position, or to be in a condition, or to do, or to undergo. "Substance," for example, "man," "horse"; -"how much," for example, "two cubits," "three cubits"; -"of what kind," for example, "white," "educated"; "relating to what," for example, "double," "half," "larger"; -"where," for example, "at the Lyceum," "at the market"; -"when," for example, "today," "last year"; -"to be in a position," for example, "he is lying down," "he is seated"; -"to be in a condition," for example, "he is shod," "he is armed"; -"to do," for example, "he cuts," "he burns"; -"to undergo," for example, "he is cut," "he is burned."

Aristotle thus posits the totality of predications that may be made about a being, and he aims to define the logical status of each one of them. Now it seems to us-and we shall try to show-that these distinctions are primarily categories of language and that, in fact, Aristotle, reasoning in the absolute, is simply identifying certain fundamental categories of the language in which he thought. Even a cursory look at the statement of the categories and the examples that illustrate them, will easily verify this interpretation, which apparently has not been proposed before. Let us consider the ten terms in order. It does not matter here if one translates ovata as "substance" or "essence." What does matter is that the category gives to the question "what?" the reply, "man" or "horse," hence the specimens of the linguistic class of nouns indicating objects, whether these are concepts or individuals. We shall come back a little later to the term ova{a to denote this predicate. The two following terms, noaov and nOLov, make a pair. They refer to 'being of what degree' [etre-quantieme < OF quant < Lat. quantus 'how great', < quam 'to what degree'], hence the abstraction, noao-r'Y}' 'quant-ity' [quant-iti

PROBLEMS IN GENERAL LINGUISTICS

'how-much-ness'], and to being of what sort [etre-quel < Lat. qualis 'of what sore], hence the abstraction, 7loton7~ 'qual-ity' [qual-iii 'what-sort-ness']. The first does not properly imply "number," which is only one of the varieties of noaov, but more generally everything capable of measurement; thus the theory distinguishes discrete "quantities" like number or language, and continuous "quantities" like straight lines or time or space. The category of notov includes the what-ness [qual-ite] without the acceptation of species. As for the three following words, neo~ il, nov, and nOTi, they refer unambiguously to 'relationship,' 'place,' and 'time.' Let us focus our attention on these six categories in their nature and in their grouping. It seems to us that these predications do not refer to attributes discovered in things, but to a classification arising from the language itself. The notion of ova{a points to the class of substantives. That of noaov and 1WlOV cited together does not only correspond to the class of adjectives in general but in particular to two types of adjectives which are closely associated in Greek. Even in the earliest texts, and before the awakening of philosophical thought, Greek joined or opposed the two adjectives noaol and no'iol, with the correlative forms o~o~ and olo~, as well as Toaos and TOtO~.2 These were formations deeply rooted in Greek, both derived from pronominal stems, and the second was productive; besides olo~, noto~, and ro'io~, there are aAAoio s and op,o 'io s . It is indeed in the system of the forms of the language that these two necessary predications were based. If we go on to 7leO~ il, behind the "relation" there is again a fundamental property of Greek adjectives, that of having a comparative (such as p,EiCov, given, in fact, as an example) which by function is a "relative" form. The two other examples, &nAaawv and fiP,lav, mark "relation" in a different way: it is the concept of "double" or "half" which is relative by definition, while it is the form of p,EtCOV which indicates "relation." As for nov 'where' and nOTi 'when', they involve the classes of spatial and temporal denominations, respectively, and here again the concepts are modelled on the characteristics of these denominations in Greek; not only are nov and nOTe linked together in the symmetry of their formation as reproduced in 0-6 {iTt, rov rou, but they are part of a class which includes still other adverbs (of the type of EX()i~, nievatv) or certain locative phrases (thus, lv AVXE{CP, lv ayoe~). It is, thus, not without reason that these categories are enumerated and grouped as they are. The first six refer all to nominal forms. Their unity is to be found in the particular nature of Greek morphology. By the same consideration, the four following also form a set: they are all from verbal categories. They are even more interesting for us since the nature of two of them does not seem to have been identified correctly. The last two are clear immediately: nOIEiv 'to do,' with the examples,

Categories of Thought and Language

59

and i-Callol 'he cuts,' 'he burns'; naaXeW 'to undergo,' with TEpVeTat, 'he is cut,' 'he is burned,' show the two categories of the active and passive, and this time the examples themselves are chosen in such a way as to emphasize the linguistic opposition. It is that morphological opposition of two "voices," present in a great number of Greek verbs, which shows through the polar concepts of nOlelV and naaXClv. But what about the first two categories, Xela()al and fX eW ? The translation does not even seem certain: some take fXeW as 'to have.' What interest could a category like "position" (XclO()at) possibly have? Is it a predication as general as "the active" or "the passive"? Is it even of the same nature? And what can be said of fXeW with examples like "he is shod," and "he is armed"? The interpreters of Aristotle seem to consider that these two categories are episodic; the philosopher only expressed them to exhaust all the predications applicable to a man. "Aristotle," says Gomperz, "imagines a man standing before him, say in the Lyceum, and passes in successive review the questions which may be put and answered about him. All the predicates which can be attached to that subject fall under one or other of the ten heads, from the supreme question, What is the object here perceived? down to such a subordinate question, dealing with mere externalities, as: What has he on? What equipment or accoutrements, e.g., shoes or weapons? ... The enumeration is intended to comprise the maximum of predicates which can be assigned to any thing or being."3 Such, as far as we can see, is the general opinion of scholars. If they are to be believed, the philosopher did not distinguish clearly between the essential and the accessory and even gave these two secondary notions precedence over a distinction like that between the active and passive. Here again, these notions seem to us to have a linguistic basis. Let us first take the Xelo()al. What could a logical category of Xela()al answer to? The answer is in the examples cited: aVa'KelTal 'he is lying down' and xa()YjTm 'he is seated.' These are two specimens of middle verbs. From the standpoint of the Greek language, that is an essential notion. Contrary to the way it appears to us, the middle voice is more important than the passive, which is derived from it. In the verbal system of ancient Greek, such as it still existed in the classic period, the real distinction was between the active and the middle. 4 A Greek thinker could with good reason set up in the absolute a predication expressed by means of a specific class of verbs, those which are only middles (the media tan tum) and mean, among other things, "position" or "attitude." Equally divided from either the active or the passive, the middle denotes a manner of being just as specific as the two others. Much the same is true of the predication called fXElV. It cannot be taken in the usual sense of fXEW 'to have,' a "having" of a material possession. What is peculiar and at first sight misleading about this category is brought to

TiflVH

'KaleTat

60

PROBLEMS IN GENERAL LINGUISTICS

light by the examples: v7robebE:rat 'he is shod' and (JbAunat 'he is armed' and Aristotle stresses this when he returns to the subject (Categories, 9); he uses the same examples a propos of EXELl', this time in the infinitive: Tel v7rooElJiaOat,'io (07r),,{aOru. The key to the interpretation is in the nature of these verbal forms; {moOiOEiat and (tm}.taTat are perfects. They are even, to speak precisely, middle perfects. But the characteristic of the middle was already expressed, as we have just seen, by y.EiaOw, whose derivatives, a'l'Uy.Errat and ;UJ.()}7iat, given as examples, do not, incidentally, have perfects. In the predication EXav and in the two forms chosen to illustrate it, it is the category of the perfect which is to the fore. The meaning of tXElJl, both 'to have' and, used absolutely, 'to be in a certain state,' agrees best with the diathesis of the perfect. \Vithout embarking upon a commentary which could easily be prolonged, let us only consider that for bringing out the value of the perfect in the translation of the cited forms, we must include in it the notion of "to have" ; they will then become, V7rOOibEiat 'he has his shoes on his feet,' W7rAWTat 'he has his armor on him.' Let us notice again that these two categories, such as we understand them, follow one another in the enumeration and seem to form a pair, just like 7Wtct'll and nuaxEt'll which follow. There are, indeed, various relationships, both formal and functional, between the Greek perfect and the middle voice, which, as inherited from Indo-European, formed a complex system; for example, an active perfect, yeyol'a goes with middle present ytY'JJOf.1 at . These relationships created many difficulties for the Greek grammarians of the Stoic school; sometimes they defined the perfect as a distinct tense, the naQaXdf.1E'IIOr; or the iiAEWr;; sometimes they set it with the middle in the class called f.1 EaoT 1]r;, intermediate between the active and the passive. Surely in any case the perfect is not easily inserted into the tense system of Greek and remains apart as indicating, as the case might be, either a mode of temporality or a manner of being in the subject. For that reason, it is understandable in view of the number of notions expressed in Greek only by the perfect, that Aristotle made it into a specific mode of being, the state (or habitus) of the subject. The ten categories can now be transcribed in linguistic terms. Each of them is given by its designation and followed by its equivalent: ova[a (,substance'), substantive; noao1J , nOlO'll (,what, in what number'), adjectives derived from pronouns like the Latin qualis and quantus; neor; it (,relating to what'), comparative adjective; nov (,where'), nOTe (,when'), adverbs of place and time; y.Ela()at ('to be placed'), middle voice; SXEW ('to be in a state'), the perfect; notEl'JJ ('to do'), active voice; naaXEt'll ('to undergo'), passive voice. In working out this table of "categories," Aristotle intended to list all the possible predications for a proposition, with the condition that each term be meaningful in isolation, not engaged in a aVf.1nAOI'C17, or, as we would say, in a

Categories of Thought and Language

61

syntagm. Unconsciously he took as a criterion the empirical necessity of a distinct expression for each of his predications. He was thus bound to reflect unconsciously the distinctions which the language itself showed among the main classes of forms, since it is through their differences that these forms and these classes have a linguistic meaning. He thought he was defining the attributes of objects but he was really setting up linguistic entities; it is the language which, thanks to its own categories, makes them to be recognized and specified. \Ve have thus an answer to the question raised in the beginning which led us to this analysis. \Ve asked ourselves what was the nature of the relationship between categories of thought and categories of language. No matter how much validity Aristotle's categories have as categories of thought, they turn out to be transposed from categories of language. It is what one can say which delimits and organizes what one can think. Language provides the fundamental configuration of the properties of things as recognized by the mind. This table of predications informs us above all about the class structure of a particular language. It follows that what Aristotle gave us as a table of general and permanent conditions is only a conceptual projection of a given linguistic state. This remark can be elaborated further. Beyond the Aristotelian terms, above that categorization, there is the notion of "being" which envelops everything. Without being a predicate itself, "being" is the condition of all predicates. All the varieties of "being-such," of "state," all the possible views of "time," etc., depend on the notion of "being." Now here again, this concept reflects a very specific linguistic quality. Greek not only possesses a verb "to be" (which is by no means a necessity in every language), but it makes very peculiar uses of this verb. It gave it a logical function, that of the copula (Aristotle himself had remarked earlier that in that function the verb did not actually signify anything, that it operated simply as a synthesis), and consequently this verb received a larger extension than any other whatever. In addition, "to be" could become, thanks to the article, a nominal notion, treated as a thing; it gave rise to varieties, for example its present participle, which itself had been made a substantive, and in several kinds (ro ov, Ot ovrcC;, Tel ovra); it could serve as a predicate itself, as in the locution ro ri 17v elVat designating the conceptual essence of a thing, not to mention the astonishing diversity of particular predicates with which it could be construed, by means of case forms and prepositions .... Listing this abundance of uses would be endless; but they really are facts of language, of syntax, and of derivation. Let us emphasize this, because it is in a linguistic situation thus characterized that the whole Greek metaphysic of "being" was able to come into existence and develop-the magnificent images of the poem of Parmenides as well as the dialectic of The Sophist. The language did not, of course, give direction to the metaphysical definition

62

PROBLEMS IN GENERAL LINGUISTICS

of Clbeing" -each Greek thinker has his own-but it made it possible to set up "being" as an objectifiable notion which philosophical thought could handle, analyze, and define just as any other concept. That this is primarily a matter of language will be better realized if the behavior of this same notion in a different language is considered. It is best to choose a language of an entirely different type to compare with the Greek, because it is precisely in the internal organization of their categories that linguistic types differ the most. Let us only state that what we are comparing here are facts of linguistic expression, not conceptual developments. In the Ewe language (spoken in Togo), which we have chosen for this contrast, the notion of Clto be," or what we shall designate as such, is divided 3mong several verbs. 5 First of all there is a verb, nye, which we would say serves to equate subject and predicate; it states, Clto be someone, to be something." The curious fact is that nye behaves like a transitive verb and governs as a complement in the accusative what for us is a predicate nominative. A second verb is Ie, which properly expresses "existence": Mawu Ie 'God exists.' But it also has a predicative use; Ie is used with predicates of situation, of localization, "to be" in a state, in a time, in a quality: e-Ie nyuie 'he is well'; e-Ie a ji 'he is here'; e-Ie /:ZO me 'he is at horne.' All spatial and temporal determination is thus expressed by Ie. Now, in all these uses, Ie exists in only one tense, the aorist, which fulfils the functions of a narrative past tense and also of a present perfect. If the predicative sentence involving Ie has to be put into another tense, like the future or the habitual, Ie is replaced by the transitive verb no 'to remain, to stay'; that is to say, depending on the tense employed, two distinctive verbs are necessary: the intransitive Ie or the transitive no, for the same notion. A verb, wo 'to accomplish, produce an effect,' with certain nouns denoting substances, behaves in the manner of our Clto be" followed by an adjective denoting substance: wo with ke 'sand,' gives wo ke 'to be sandy'; with tsi 'water,' wo tsi 'to be wet'; with kpe 'stone,' wo kpe 'to be stony.' What we take as a "being" by nature is in Ewe a "making," like the French "ilfait du vent." When the predicate is a term of function or of rank, the verb is du, hence du jia 'to be king.' Finally, with certain predicates of physical quality or of state, "to be" is expressed by di; for example, di ku 'to be thin'; di fo 'to be a debtor.' In practice there are thus five distinct verbs which correspond approximately to the functions of our verb "to be." This does not mean that the same semantic area is divided into five portions; it is a distribution which brings about a different arrangement, even extending into neighboring notions. For instance, the two notions of "to be" and "to have" are as distinct for us as the terms that

Categories of Thought and Language

express them. Now, in Ewe, one of the verbs cited, Ie, the verb of existence, when joined to asi 'in the hand,' forms the locution Ie asi, literally, 'to be in the hand,' which is the most usual equivalent for our "to have"; ga Ie asi-nye (literally, 'money is in my hand') 'I have money.' This description of the state of things in Ewe is a bit contrived. It is made from the standpoint of our language and not, as it should have been, within the framework of the language itself. Within the morphology or syntax of Ewe, nothing brings these five verbs into relationship with one another. It is in connection with our own linguistic usages that we discover something common to them. But that is precisely the advantage of this "egocentric" comparison: it throws light on ourselves; it shows us, among that variety of uses of "to be" in Greek, a phenomenon peculiar to the Indo-European languages which is not at all a universal situation or a necessary condition. Of course the Greek thinkers in their turn acted upon the language, enriched the meanings, and created new forms. It is indeed from philosophical reflection on "being" that the abstract substantive derived from elVat arose; we see it being created in the course of history: at first as laa{a in Dorian Pythagorism and in Plato, then as ova{a, which won out. All we wish to show here is that the linguistic structure of Greek predisposed the notion of "being" to a philosophical vocation. By comparison, the Ewe language offers us only a narrow notion and particularized uses. We cannot say what place "being" holds in Ewe metaphysics, but, a priori, the notion must be articulated in a completely different way. It is the nature of language to give rise to two illusions of opposite meaning: being learnable, consisting of an always limited number of elements, language gives the impression of being only one of the interpreters possible for thought, while thought, being free, autarchical, and individual, uses language as its instrument. As a matter of fact, whoever tries to grasp the proper framework of thought encounters only the categories of language. The other illusion is the opposite. The fact that language is an ordered totality and that it reveals a plan, prompts one to look in the formal system of language for the reflection of a "logic" presumably inherent in the mind and hence exterior and anterior to language. By doing this, however, one only constructs naIvetes or tautologies. Surely it is not by chance that modern epistemology does not try to set up a table of categories. It is more productive to conceive of the mind as a virtuality than as a framework, as a dynamism than as a structure. It is a fact that, to satisfy the requirements of scientific methods, thought everywhere adopts the same procedures in whatever language it chooses to describe experience. In this sense, it becomes independent, not of language, but of particular linguistic structures. Chinese thought may well have invented categories as specific as the tao, the yin, and the yang; it is nonetheless able to assimilate the concepts

PROBLEMS IN GENERAL LINGUISTI

of dialectical materialism or quantum mechanics without the structure of t: Chinese language proving a hindrance. No type of language can by itself alo: foster or hamper the activity of the mind. The advance of thought is link, much more closely to the capacities of men, to general conditions of cultUl and to the organization of society than to the particular nature of a langua~ But the possibility of thought is linked to the faculty of speech, for language a structure informed with signification, and to think is to manipulate the sig of language.

From Les Etudes philosophiques, no. 4 (October-December 1958), pp. 419-429

SEVEN

Remarks on the Function of Language in Freudian Theory

To THE DEGREE THAT psychoanalysis intends to establish itself as a science, one is justified in challenging its method, its procedures, and its purpose, and in comparing them with those of the recognized "sciences." Whoever wants to make out the ways of reasoning upon which the analytic method rests is brought to a remarkable observation. From the moment a disturbance is observed until its cure, it looks as though nothing material were in operation. Nothing is done which lends itself to an objective observation. That visible causal relationship which one looks for in scientific reasoning does not appear between one induction and another. When, in contrast to the psychoanalyst, the psychiatrist attempts to relate the disturbance to an injury, at least his procedure has the classic appearance of an investigation which goes back to the "cause" in order to treat it. There is nothing like this in the analytical technique. For the person who knows analysis only from the accounts which Freud l has given of it (and this is the case of the author of these pages) and who considers less the practical efficacy, which is not in question here, than the nature of the phenomena and the relationships in which they are set, psychoanalysis seems to be different from all other disciplines. The principal difference is this: the analyst operates on what the subject says to him. He considers him in the discourses which he holds with him and examines him in his locutory or "fabulatory" behavior, and through the patient's discourses another discourse slowly takes shape for the analyst, one which he will endeavor to explain: that of the complex buried in the unconscious. The success of the cure depends on bringing this complex to light, and this in turn testifies to the correctness of the induction. Thus, from patient to analyst, and from analyst to patient, the entire process operates through language. It is this relationship which deserves notice and properly marks this type of analysis. It brings out, it seems to us, that all of the diverse symptoms which the analyst encounters and scrutinizes in succession are the product of an initial motivation in the patient that is unconscious to the utmost and is often transposed into other motivations, conscious and usually misleading. Starting

66

PROBLEMS IN GENERAL LINGUISTICS

from this motivation, which has to be brought to light, all the actions of the patient are illuminated by and linked to the disturbance which, in the eyes of the analyst, is both its manifestation and the symbolic substitute for it. \Ve thus perceive here an essential trait of the analytical method; the "phenomena" are governed by a moti'vational relationship which here holds the place of what the natural sciences define as a causal relationship. It seems to us that if analysts acknowledge this view, the scientific status of their discipline, with its special individuality, and the specific character of their method would be better established. There is a clear indication that in this field, motivation functions as "cause." As we know, the procedure of the analyst is entirely regressive and aims to provoke the emergence, in the memory and in the discourse of the patient, of the actual datum around which the analytic unravelling of the morbid behavior will aftenvard center. The analyst is thus in search of a "historical" datum which lies hidd~n and unknown in the memory of the subject, whether or not the latter consents to "recognize" it and identify himself with it. It might be objected that this bringing to light again of an actual fact, of a biographical experience, is really equivalent to the discovery of a "cause." But one sees immediately that the biographical fact cannot bear the burden of the causal connection all by itself, chiefly because the analyst cannot recognize it without the aid of the patient, who is the only one to know "what happened to him." Even if he could, he would not know what value to assign to the fact. Let us even suppose that in a Utopian universe, the analyst could discover, from objective evidence, the traces of all the events that make up the biography of the patient; still he would not be able to infer very much from this, not even what is essential, unless by a happy chance. For if he needs the patient to tell him everything and even to express himself at random and aimlessly, it is not in order to recover an empirical fact, which will not have been registered at all except in the patient's memory; it is because empirical facts have no reality for the analyst except in and through the "discourse" which gives them the authenticity of an actual experience, without regard to their historical reality and even (perhaps, especially) if the discourse evades, transposes, or invents the biography which the patient gives himself. This is the case for the very reason that the analyst wishes to unveil motivations rather than identify events. The constitutive dimension of that biography is that it is verbalized and thus assumed by the one who is telling about himself. Its expression is that of language, and the relationship of the analyst to the subject is that of dialogue. Everything here proclaims the advent of a technique that makes human language its field of action and the special instrument of its efficacy. But then a fundamental question arises: just what is this "language" which acts as much as it expresses something? Is it identical with that which one uses out-

Language in Freudian The01Y side of analysis? Is it even the same for the two partners? In his brilliant treatise on the function and field of the individual act of sp~aking [parole] and language [langage] in psychoanalysis, Lacan said about the analytic method (p. 103): 'Its means arc those of parole in that it confers a meaning upon the functions, of an individual; its domain is that of concrde discourse as the transindividual reality of the subject; its operations arc historical in that it constitutes the emergence of the truth into reality."2 Starting from these precise definitions, and primarily from the distinction brought up between the means and the domain, we may attempt to delimit the varieties of "language" which are involved. In the first instance, we recognize the universe of the individual act of speech [parole], \\"hich is that of subjectivity. All through Freudian analysis it can be seen that the suhject makes use of the act of speech and discourse in order to "represent himself" to himself as he wishes to see himself and as he calls upon the "other" to observe him. His discourse is appeal and recourse: a sometimes vehement solicitation of the other through the discourse in which he figures himself desperately, and an often mendacious recourse to the other in order to individualize himself in his own eyes. Through the sole fact of addressing another, the one who is speaking of himself installs the other in himself, and thereby apprehends himself, confronts himself, and establishes himself as he aspires to be, and finally historicizes himself in this incomplete or falsified history. Language [langage] is thus used here as the act of speech [parole], converted into that expression of instantaneous and elusive subjectivity which forms the condition of dialogue. The subject's language [langue] provides the instrument of a discourse in which his personality is released and creates itself, reaches out to the other and makes itself be recognized by him. Now a language [langue] is a socialized structure which the act of speaking [parole] subjects to individual and intersubjective ends, thus adding to it a new and strictly personal design. Langue is a system common to everyone; discourse is both the bearer of a message and the instrument of action. In this sense, the configurations of every act of speaking are unique, realized within and by means of language. There is thus an antinomy within the subject between discourse and language. But for the analyst, the antinomy establishes itself on a very different plane and assumes another meaning. He must be attentive to the content of the discourse, but no less and especially to the gaps in the discourse. If the content informs him about the image which the subject has of the situation and about the position in it that he attributes to himself, he searches through this content for a new content: that of the unconscious motivation that proceeds from the buried complex. Beyond the innate symbolism of language, he will perceive a specific symbolism which will be formed, without the subject

68

PROBLEMS IN GENERAL LINGUISTICS

being aware of it, as much from what is omitted as from what is stated. And within the history in which the subject places himself, the analyst will provoke the emergence of another history, which will explain the motivation. He will thus take the discourse as the translation of another "language," which has its own rules, symbols, and "syntax," and which goes back to the deep structures of the psyche. These distinctions would require many developments that the analyst alone could state with precision and with attention to all the shades of difference. In marking them we should like above all to clear up certain confusions that could easily become fixed in a field in which it is already difficult to know what one means when investigating "naive" language, and where analytical concern creates a new difficulty. Freud cast significant light upon verbal activity as revealed in its lapses, in its play aspects, and in its free wanderings when the power of repression is suspended. All the anarchical force that is repressed or sublimated in normal language is rooted in the unconscious. Freud also noticed the deep affinity between these forms of language and the nature of the associations that are made in dreams (another expression of unconscious motivations). He was thus led to reflect upon the functioning of language in its connections with the subconscious structures of the psyche and to wonder if the conflicts which defined this psyche had not left their imprint upon the very forms of language. Freud posed the problem in an article published in 1910 entitled, "The Antithetical Sense of Primal Words."3 As a point of departure there was an essential observation from his Traumdeutung on the insensitivity to contradiction that characterizes the logic of dreams: The attitude of dreams towards the category of antithesis and contradiction is most striking. This category is simply ignored; the word "No" does not seem to exist for a dream. Dreams show a special tendency to reduce two opposites to a unity or to represent them as one thing. Dreams even take the liberty, moreover, of representing any element whatever by the opposite wish, so that it is at first impossible to ascertain, in regard to any element capable of an opposite, whether it is to be taken negatively or positively in the dream-thoughts. Now Freud believed that he had found in a study by Karl Abel the proof that "this habit of the dream-work to which I refer exactly tallies with a peculiarity in the oldest languages known to us." After giving some examples, he concluded: In the agreement between that peculiarity of the dreamwork mentioned at the beginning of this paper and that which philologists have discovered to be habitual in the oldest languages, we may see a confirmation of our supposi-

Language in Freudian Theory tion in regard to the regressive, archaic character of thought-expression in dreams. And we cannot dismiss the conjecture, which forces itself on us psychiatrists, that we should understand the language of dreams better and translate it more easily if we knew more about the development of language.

The authority of Freud is almost enough to validate this demonstration or at least give sanction to the idea that it might lead to fruitful research. An analogy would seem to have been discovered between the workings of dreams and the semantics of "primitive" languages in which the same term states one thing and its opposite as well. The way would seem open to an investigation that would probe into the structures common to the collective language and the individual psyche. In the face of such prospects, it might be useful to state that reasons of fact deny the credibility of the etymological speculations of Karl Abel that intrigued Freud. Here we have to do no longer with psychopathological manifestations of language but with the concrete, general, and verifiable data of historical languages. It is no accident that no qualified linguist, either at the time at which Abel was writing (and there were some as early as 1884) or since, supported the methods or the conclusions of the Gegensinn der Urworte. If one claims to trace the course of the semantic history of words and to reconstruct their prehistory, the first methodological principle is to consider the forms and meanings as attested successively at each period of history all the way back to the earliest date, and to start reconstruction from the most remote point attainable in the investigation. This brings up another principle of the comparative technique, which is to submit the comparisons between languages to regular .correspondences. Abel worked without regard for these rules and assembled all the data on the basis of resemblance. From a resemblance between a German word and an English or Latin one of a different or opposite meaning, he concluded that there was a primal relationship between them by means of "opposite meanings," and he neglected all the intermediary stages which would account for the divergence when there really was a relationship or which would rule out the possibility of a relationship by proving that the words were of different origins. It is easy to show that none of the evidence brought forward by Abel could be maintained. In order not to prolong this discussion, we will restrict ourselves to examples taken from western languages which might puzzle readers who are not linguists. Abel gives a series of correspondences between English and German which Freud quoted as showing how a word in one language can have a meaning opposite to that of its cognate in another, and among which one could presumably observe "a phonetic transformation with the aim of separating the contraries." Without stressing for the moment the grave error in reasoning

PROBLEMS IN GENERAL LINGUISTICS

hidden behind that simple remark, let us be content with correcting those comparisons. The Old German adverb bass 'well' is related to besser but has no connection with bas 'bad,' just ac; in Old English, bat 'good, better' is not related to badde (modern 'bad'). English cleave does not correspond to German kleben 'to stick,' as Abel says, but to klieben 'to cleave' (cf. Kluft). English lock is not opposite to German Lucke, Loch; it is just the other way around, for the former sense of Loch is 'entrenchment, enclosed and hidden place.' German stumm properly means 'paralyzed (in the tongue)' and is connected to stammeln, stemmen, and has nothing in common with Stimme, which means 'voice' in its most ancient form, Gothic stibna. In the same way, in Latin, clam 'secretly' is linked to celare 'to hide,' and not at all to clamare, and so on. A second series of proofs, also erroneous, was drawn by Abel from certain expressions that have opposite meanings within the same language. Such would be the double sense of Latin sacer 'consecrated' and 'accursed.' Here the ambivalence of the notion should no longer surprise us since so many studies of the phenomenology of the sacred have made a cliche of its basic duality; in the Middle Ages, a king and a leper were both "untouchables," but it does not follow that sacer includes two contradictory senses; it was cultural conditions which determined two opposed attitudes toward the object described as sacer. The double meaning attributed to Latin altus, 'high' and 'deep,' stems from the illusion that makes us take the categories of our own language as necessary and universal. Even in French one speaks of "the depths" [Ia profondeur] of the sky and of "the depths" of the sea. More precisely, the direction designated by altus is measured in an upward direction, e.g., upwards from the bottom of a well or upwards from the foot of a tree, without regard to the position of the observer, while in French, profond ['deep'] is defined in opposite directions, starting with the observer and going towards the farthest limit [Ie fond], be it that of a well or of the sky. There is nothing "primal" about these various linguistic constructions of our perceptions. Furthermore, it is not "in the origins of language" that one should search for the explanation of English without but simply in the beginnings of English. Contrary to what Abel believed and to what some people still believe, without does not include the contradictory expressions "with" and "lacking"; the proper sense of with here is "against" (cf. withstand) and indicates motion or effort in some direction. Hence, within 'toward the interior' and without 'toward the exterior,' from which comes 'outside, lacking.' To understand how German wider means 'against' and wieder (with a simple graphic variation), 'again,' it is enough to think of the same apparent contrast of the re- in French between re-pousser ['to push back'] and re-venir ['to come back']. There is no mystery in any of this, and the application of elementary rules will dissipate these mirages.

Language in Freudian Theory

71

But with that, the possibility of an homology between the stages of a dream and the processes of "primitive languages" vanishes. The question here has two aspects. One concerns the "logic" of language. Insofar as it is a collective and traditional institution, every language has its anomalies and inconsistencies, which express a dissymmetry inherent in the nature of the linguistic sign. But it nevertheless remains that language is a system, that it obeys a specific plan, and that it is articulated by a set of relationships capable of a certain formalization. The slow but incessant work that goes on within a language does not operate at random; it bears on those relationships or oppositions which are or are not necessary, in such a way as to renew or multiply distinctions that are useful at all levels of expression. The semantic organization of a language shares in this systematic character. Language is the instrument by which the world and society are adjusted; it operates on a world considered to be "real" and reflects a "real" world. But in this each language is specific and shapes the world in its own way. The distinctions each language brings forth must be explained by the particular logic that supports them and not be submitted straight off to a universal evaluation. In this regard, ancient or archaic languages are neither more nor less strange than those we speak; they only have that strangeness which we attribute to unfamiliar objects. Their categories, oriented differently from ours, are nonetheless consistent. It is thus a priori improbable-and an attentive examination confirms it-that these languages, however archaic they are assumed to be, escape the "principle of contradiction" by using the same expression for two mutually exclusive or simply contrary notions. In actual fact, we are still waiting to see serious examples of this. Let us suppose that a language exists in which "large" and "small" are expressed identically; this would be a language in which the distinction between "large" and "small" literally has no meaning and in which the category of dimension does not exist, and not a language allowing for a contradictory expression of dimension. The claim that the distinction exists but that it is not verbalized would demonstrate the insensitivity to contradiction not in the language but in the researcher, for it is indeed contradictoriness to impute to a language both a knowledge of two notions as opposite and the expression of these notions as identical. I t is the same with the logic of dreams. If we characterize the unfolding of a dream by the total freedom of its associations and by the impossibility of acknowledging an impossibility, it is primarily because we retrace and analyze it within the framework of language, and the quality of language is to express only what it is possible to express. This is not a tautology. A language is primarily a categorization, a creation of objects and of relations between those objects. To imagine a stage of language as "primal" as one would wish, but nevertheless real and "historical," in which a certain object would be

PROBLEMS IN GENERAL LINGUISTICS

denominated as being itself and at the same time something else, and in which the relation expressed would be a relation of permanent contradiction-a nonrelating relation-in which everything would be itself and something else, hence neither itself nor the other, is to imagine a pure chimera. Insofar as we can avail ourselves of the evidence of "primitive" languages to go back to the origins of linguistic experience, we have instead to envisage an extreme complexity of classification and a multiplicity of categories. Everything seems to take us far away from an "actual" correlation between oneiric logic and the logic of a real language. Let us note also in passing that, as it happens in "primitive" societies, far from the language reproducing the appearance of a dream, it is the dream which is brought to the categories of the language. The dream is interpreted in connection with actual situations and by means of a set of equivalences that submit it to a real linguistic rationalization. 4 What Freud asked in vain of "historical" language he could have asked to a certain extent of myth or poetry. Certain forms of poetry can be related to dreams and can suggest the same mode of structuring. They can bring about in the normal forms of language that suspension of meaning which dreams project into our activities. But that means, paradoxically, that Freud could have found in surrealist poetry (which, according to Breton, he did not understand) something of what he was seeking, wrongly, in organized language. These confusions seem to have arisen in Freud from his constant recourse to "origins": origins of art, of religion, of society, of language .... He was constantly transposing what seemed to him to be "primitive" in man into an original primitivism, for it was indeed into the history of this world that he projected what we could call a chronology of the human psyche. Was that legitimate? \Vhat ontogenesis allows the analyst to set up as archetypal is only so with respect to what distorts it or represses it. But if one makes of this repression something which is genetically coextensive with society, one can no more imagine a society without conflict than a conflict outside of society. R6heim discovered the Oedipus complex among the most "primitive" societies. If this complex is inherent in society as such, an Oedipus who is free to marry his mother is a contradiction in terms. And in this case, it is nothing other than conflict which is nuclear in the human psyche. But then the notion of "primal" no longer makes sense. As soon as one posits organized language in correspondence to the elementary psyche, one introduces into the argument a new datum which breaks the symmetry that one had thought to establish. Freud gave the proof of this himself without knowing it in his ingenious essay on negation. 5 He reduced the polarity of linguistic affirmation and negation to the biopsychical mechanism of acceptance within oneself or rejection outside oneself, connected with

Language in Freudian Theory

73

the appreciation of good and evil. But animals are also capable of this evaluation which leads to acceptance in the self and rejection outside the self. The characteristic of linguistic negation is that it can annul only what has been uttered, which it has to set up for the express purpose of suppressing, and that a judgment of nonexistence has necessarily the formal status of a judgment of existence. Thus negation is first acceptance. Completely different is the preliminary refusal to accept, which is called repression. Freud himself stated very well what negation shows: Thus the content of a repressed image or idea can make its way into consciousness, on condition that it is negated. Negation is a way of taking cognizance of what is repressed; indeed it is already a lifting of the repression, though not, of course, an acceptance of what is repressed .... The outcome of this is a kind of intellectual acceptance of the repressed, while at the same time what is essential to the repression persists.

Is it not apparent here that the linguistic factor is decisive in this complex process, and that negation is in some way constitutive of the denied content, hence of the emergence of this content in the consciousness and the suppression of the repression? What then survives of the repression is only a repugnance to be identified with this content, but the subject no longer has power over the existence of this content. Here again, his discourse can produce denials in abundance but it cannot abolish the fundamental property of language, which is to imply that something corresponds to what is uttered, some thing and not "nothing." We reach with this the essential problem, the immediacy of which all these discussions and the ensemble of analytic procedures bear witness to: that of symbolism. All psychoanalysis is grounded on a theory of symbolism. Now, language is nothing but symbolism. But the differences between the two symbolisms illustrate and sum up all the differences that we have been stressing all along. The profound analyses that Freud gave to the symbolism of the unconscious also illuminate the different ways by which the symbolism of language is realized. In saying that language is symbolic, one states only its most manifest property. It must be added that language is necessarily realized in a specific language, and thus a difference appears which defines linguistic symbolism for man: it is learned; it is coextensive with man's acquisition of the world and of intelligence, with both of which he finally becomes unified. It follows that for man the main symbols and their syntax cannot be separated from things and from the experience he has of them; he must master them in proportion as he discovers them as realities. Whoever comprehends the diversity of these symbols as actualized in the various terms of various languages soon realizes that the relationship of these symbols to the

74

PROBLEMS IN GENERAL LINGUISTICS

things that seem to underlie them can only be acknowledged, not justified. In comparison with this symbolism which is expressed in infinitely varied signs, combined in formal systems as numerous and distinct as there are languages, the symbolism of the unconscious discovered by Freud shows absolutely specific and different characteristics. Some of these should be stressed. First, its universality. It seems, according to the studies made of dreams or neuroses, that the symbols that translate them constitute a "vocabulary" common to all peoples without respect to language. This is due to the fact, of course, that they are not learned or recognized as language by those who produce them. Furthermore, the relationship between these symbols and what they represent can be defined by the abundance of the signifiers and the uniqueness of the signified; this stems from the fact that the content is repressed and can be released only in the guise of images. On the other hand, in contrast to the linguistic sign, these multiple signifiers and this unique signified are constantly linked by a "motivational" connection. Let us observe, finally, that the "syntax" in which these unconscious symbols are strung together does not obey any logical necessity, or rather, that it recognizes only one dimension, that of succession, which, as Freud saw, also signifies causality. We are thus confronted with a "language" so special that it is of the greatest importance to distinguish it from what we normally call language. It is by stressing the differences that one can better locate it in the catalogue of linguistic expressions. "This symbolism," Freud said, "is not peculiar to dreams; but is characteristic of unconscious ideation, in particular among the people, and it is to be found in folklore, and in popular myths, legends, linguistic idioms, proverbial wisdom, and current jokes to a more complete extent than in dreams."6 This sets exactly the level of the phenomenon. In the area in which this unconscious symbolism appears, one could say that it is both infra- and supralinguistic. As infralinguistic, it has its source in a region deeper down than that in which education installs the linguistic mechanism. It makes use of signs that cannot be split up and that admit of numerous individual variants, susceptible themselves of being increased by reference to the common domain of a culture or to personal experience. It is supralinguistic in that it makes use of extremely condensed signs which, in organized language, would correspond more to large units of discourse than to minimal units. And a dynamic relationship of intentionality is established among these signs that amounts to a constant motivation (the "realization of a repressed desire") and that follows the most remarkably indirect paths. We thus come back to "discourse." By following this comparison, one would be put on the way to productive comparisons between the symbolism of the unconscious and certain typical procedures of the subjectivity manifested in discourse. On the level of speech, one can be precise: these are the

Language in Freudian Theory

75

stylistic devices of discourse. For it is style rather than language that we would take as term of comparison with the properties that Freud has disclosed as indicative of oneiric "language." One is struck by the analogies which suggest themselves here. The unconscious uses a veritable "rhetoric" which, like style, has its "figures/' and the old catalogue of tropes would supply an inventory appropriate to the two types of expression. One finds in both all the devices of substitution engendered by taboo: euphemism, allusion, antiphrasis, preterition, litotes. The nature of the content makes all the varieties of metaphor appear, for symbols of the unconscious take both their meaning and their difficulty from metaphoric conversion. They also employ what traditional rhetoric calls metonymy (the container for the contents) and synechdoche (the part for the whole), and if the "syntax" of the symbolic sequences calls forth one device of style more than any other, it is ellipsis. In short, to the extent that symbolic images in myths and dreams, etc., will be listed, one will probably see more clearly into the dynamic structures of style and their affective components. What is intentional in motivation obscurely controls the manner in which the inventor of a style fashions common material and, in his own way, releases himself therein. For what is called unconscious is responsible for the way in which the individual constructs his persona, and for what he accepts and what he rejects or fails to recognize, the former being motivated by the latter.

From La Psych analyse

I

(1956) : 3-16

Structures and Analyses

Structures and Analyses

EIGHT

"Structure" in Linguistics

IN THE COURSE OF THE last twenty years, the term "structure" has undergone considerable extension after acquiring a doctrinal and somewhat programmatic value. "Moreover, it is not structure that seems to be the essential term now so much as the adjective structural, used to qualify a kind of linguistics. Structural rapidly led to structuraHsm and structuralist. Thus an ensemble of designations 1 was created which other disciplines have now borrowed from linguistics in order to adapt them to their own values. 2 Today one cannot glance over the table of contents of a linguistics journal without meeting one of these terms, often in the title of the work. We will readily admit that a concern for being "modern" is not foreign to this widespread use and that certain "structuralist" pronouncements cover works whose novelty or interest is debatable. The object of the present note is not to denounce the abuse but to explain the use. The question is not to assign its field and limits to "structural" linguistics but to clarify what is involved in the concern with structure and what meaning this term had for those linguists who first gave it a precise meaning. 3 The principle of "structure" as a topic for study was asserted a little before 1930 by a small group of linguists who proposed to react thus against the exclusively historical concept of language, against a linguistics that broke language down into isolated elements and was engaged in following the changes that took place in them. It is agreed that this movement had its origin in the teachings of Ferdinand de Saussure at Geneva, as they were put into writing by his students and published under the title Cours de linguistique generale. 4 Saussure is rightly called the precursor of modern structuralism. s He certainly was, except for the term. It is important to note, for exactitude in describing this movement of ideas which must not be simplified, that Saussure never used the word "structure" in any sense whatever. In his eyes, the essential notion was system. In that was the novelty of his doctrine, in the idea -so full of implications that it took a long time to perceive and develop-that language forms a system. That is the way the Cours presented it, in statements

80

PROBLEMS IN GENERAL LINGUISTICS

that should be recalled: "Language is a system that has its own arrangement" (p. 43 [po 22]); " ... language is a system of arbitrary signs" (p. 106 [po 73]); "Language is a system whose parts can and must all be considered in their synchronic solidc.rity" (p. 124 [po 87]). And above everything else, Saussure stated the primacy of the system over the clements which composed it: " ... to consider a term as simply the union of a certain sound with a certain concept is grossly misleading. To define it in this way would isolate the term from its system; it would mean assuming that one can start from the terms and construct the system by adding them together when, on the contrary, it is from the interdependent whole that one must start and through analysis obtain its elements" (p. 157 [po 113]). This last sentence contains the germ of all that is essential in the "structural" concept. But it was always to the system that Saussure referred. This notion was familiar to Saussure's students in Paris.6 'VeIl before it was worked out in the Cours de linguistique generale, Meillet had stated it several times, without failing to ascribe it to the teaching of his master, of whom he said, "throughout his whole life what he was trying to determine was the system of the languages he was studying."7 When Meillet said that "each language is a rigorously organized system in which everything holds together,"8 it was to give Saussure the credit for having shown this in the system of Indo-EuropeaIl vowels. He came back to this several times: "It is never legitimate to explain a detail except in the context of the general system of the language in which it appears";9 "A language constitutes a complex system of means of expression, a system in which everything holds together .... "10 In the same way, Grammont praised Saussure for having shown "that each language forms a system in which everything holds together, in which the facts and phenomena control one another and can be neither isolated nor contradictory."ll In discussing "phonetic laws" he stated, "There are no isolated phonetic changes .... The whole set of articulations in a language in effect constitutes a system in which everything holds together, in which everything depends strictly on everything else. As a result, if a modification is produced in one part of the system, there is a good chance that the whole system will be affected, for it is necessary that the system remain coherent."12 Thus the notion of language as system was accepted long ago by those who were taught by Saussure, first in comparative grammar and then in general linguistics. 13 If one adds to this two other principles which are equally Saussurian: that language is form, not substance, and that the units of language can only be defined by their relationships, one will have indicated the fundamentals of the doctrine which some years later was to show the structure of linguistic systems. This doctrine was first expressed in the proposals for studying phonemic

" Structure" in Linguistics

81

systems,14 drawn up in French by three Russian linguists, R. Jakobson, S. Karcevski, and N. Trubetskoy, and addressed to the First International Congress of Linguists at The Hague in 1928. These innovators were themselves to name those whom they considered their predecessors, Saussure on the one hand, and Baudouin de Courtenay on the other. But even then their ideas had taken an autonomous form, and as early as 1929 they formulated them in the theses published in French at Prague for the First Congress of Slavic Philologists. 15 These anonymous theses, which constituted a veritable manifesto, inaugurated the activity of the Linguistic Circle of Prague. It was in them that the term structure appeared, with the value which several examples will illustrate. The title speaks of "problems of methodology stemming from the conception of language as a system," and the subtitle, " ... structural comparison and genetic comparison." They advocated "a method suitable for permitting the discovery of the laws of structure of linguistic systems and their evolution."16 The notion of "structure" was closely linked with that of "relationship" within the system: "The sensory content of phonological elements is less essential than their reciprocal relationships within the system (structural principle of the phonological system)."17 Hence this rule of method: "The phonological system must be characterized ... by an obligatory specification of the relationships existing among the said phonemes; that is, by tracing the structural scheme of the language being considered."18 These principles are applicable to all parts of the language, even to "categories of words, a system whose extent, precision, and internal structure (reciprocal relationships of its elements) must be studied for each language in particular."19 "One cannot determine the place of a word in a lexical system until one has studied the structure of the said system."20 In the collection containing these theses, several other articles by Czech linguists (Mathesius, Havranek), also written in French, contain the word "structure."21 It will be noted in the most explicit of these quotations that "structure" is complemented by the phrase "of a system." Such indeed is the sense of the term when Trubetskoy used it again a little later in an article in French on phonology:22 "To define a phoneme is to indicate its place in the phonological system, which is impossible unless one takes into account the structure of the system .... Phonology, which is universalist in nature, starts with the system as with an organic whole whose structure it studies."23 It follows that several systems can and should be confronted: "In applying the principles of phonology to several completely different languages in order to show their phonological systems, and in studying the structure of these systems, one soon perceives that certain combinations of correlations recur in the most diverse languages, while others do not exist anywhere at all. These are laws of structure of phonological systems."24 "A phonological system is not the mechanical

82

PROBLEMS IN GENERAL LINGUISTICS

sum of isolated phonemes but an organized whole of which the phonemes are the members and whose structure is subject to laws."25 According to this, the development of phonology is in accord with that of the natural sciences: "Present-day phonology is characterized above all by its structuralism and by its systematic universalism .... The age in which we live is characterized by the tendency in all the scientific disciplines to replace atomism by structuralism and individualism by universalism (in the philosophical sense of these terms, of course). This tendency can be observed in physics, in chemistry, in biology, in psychology, in economics, etc. Present-day phonology is thus not isolated. It has a place in a broader scientific movement."26 Granting that language is system, it is then a matter of analyzing its structure. Each system, being formed of units that mutually affect one another, is distinguished froIll other systems by the internal arrangements of these units, an arrangement which constitutes its structure. 27 Certain combinations are frequent, others fairly rare, and still others, while theoretically possible, are never realized. To envisage a language (or each part of a language, such as its phonetics, morphology, etc.) as a system organized by a structure to be revealed and described is to adopt the "structuralist" point of view. 28 The views of these first phonologists, based on precise descriptions of various phonological systems, gained many adherents within a few years, even outside the Linguistic Circle of Prague, so that it became possible in 1939 to found a journal in Copenhagen, Acta Linguistica, which is entitled Revue internationale de linguistique structurale. In the preliminary announcement, written in French, the Danish linguist Viggo Br0ndal justified the orientation of the journal by the importance "structure" had acquired in linguistics. In this connection he referred to the definition of the word "structure" by Lalande: "To designate, in opposition to a simple combination of elements, a whole formed of mutually dependent elements, such that each depends on the others and can only be what it is by its relationship with them."29 He also stressed the parallelism between structural linguistics and "Gestalt" psychology, by invoking the definition of the "Gestalttheorie" given by Claparede :30 "It conceives of phenomena not as a sum of elements which it is of special concern to isolate, analyze, and dissect, but as ensembles (Zusammenhange) consisting of autonomous units, manifesting an internal cohesiveness, and having their own laws. Hence the mode of being of each element depends on the structure of the ensemble and the laws which govern it."31 When Louis Hjelmslev took up the editorship of Acta Linguistica in 1944, after the death of V. Br0ndal, he defined again the domain of structural linguistics: "By structural linguistics is understood an ensemble of investigations resting on a hypothesis according to which it is scientifically legitimate to describe language as being essentially an autonomous entity of internal de-

"Structure" in Linguistics pel1dences, or, in a \vord, a structure. The analysis of this entity always allows for the disengagement of parts which affect one another reciprocally and \vhich each depends on certain others and would be neither conceivable nor definable without those other parts. This analysis sees its subject matter as a network of dependences, considering linguistic phenomena as related to one another."32 Such were the origins of "structure" and "structural" as technical terms. Today the very development of linguistic studies 33 tends to split "structuralism" into such diverse interpretations that one of those who claim allegiance to this doctrine does not hesitate to write that "under the common and misleading label of 'structuralism' are to be found schools of extremely divergent inspiration and tendencies .... The quite general use of terms like 'phoneme' and even 'structure' often serves to camouflage profound differenccs."34 One of these differences, undoubtedly the most notable, is the one which may be observed between the American use of the term "structure" and the definitions given above. 35 To limit ourselves to the use generally made of the word "structure" in European linguistics in works written in French, we shall stress some features which are capable of constituting a minimal definition of it. The fundamental principle is that a language constitutes a system whose parts are all united in a relationship of solidarity and dependence. This system organizes units, which are the articulated signs, mutually differentiating and delimiting themselves. The structuralist doctrine teaches the predominance of the system over the elements, and aims to define the structure of the system through the relationships among the elements, in the spoken chain as well as in formal paradigms, and shows the organic character of the changes to which language is subject.

From Sens et usages du terme "structure" dans les sciences humaines et sociales, ed. R. Bastide (The Hague: Mouton & Co., 1962), pp. 31-39

NINE

The Classification of Languages

A LECTURE ON A SUBJECT that would require a whole book in order to be presented and discussed in a manner adequate to its importance can claim neither to cover all the issues nor to found a new method. We are simply proposing to pass in review theories that prevail today, to show what principles underlie them and what results they can obtain. The general problem of the classification of languages can be broken down into a certain number of particular problems which vary in nature according to the type of classification envisaged. But these problems have in common the fact that each of them, formulated with rigor, challenges both the totality of the classification and the totality of the language to be classified. This is enough to make one realize the importance of the undertaking, the difficulties inherent in it, and also the distance that will appear between the goal aimed at and the means at our disposal for attaining it. The first classification tried by linguists has been that which classifies languages into families supposed to have issued from a common prototype. This is the genetic classification. The first attempts came in the Renaissance, when printing made it possible to know the languages of neighboring or distant peoples. The observations of the resemblances among these languages led rapidly to grouping them into families, which were less numerous than the existing languages, and whose differences were explained by reference to myths. With the discovery of Sanskrit and the beginnings of comparative grammar, the method of classification became rational and, without entirely abandoning the idea of the monogenesis of languages, defined with increasing precision the conditions that must be met by the establishment of a genetic relationship. Today linguists have extended procedures established by the analysis of Indo-European languages to the whole body of languages. They have grouped the majority of idioms into genetic classes. A work describing the languages of the world can hardly find a framework other than this. And while abandoning every glottogenetic hypothesis, and estimating better the limits of the knowable and the demonstrable, linguists have not, for all that,

86

PROBLEMS IN GENERAL LINGUISTICS

abandoned the search for the relationships among languages spoken in countries not yet thoroughly explored, like those of South America, for instance, or the attempt to group whole families like Indo-European and Semitic into larger units. Thus it was not the science of languages which permitted the establishment of bases for classification; on the contrary, it was from a classification, however naIve and confused, that the science of languages was gradually developed. The resemblances observed among ancient or modern European languages have been the primary data which have led to a theory of these resemblances. This observation accounts, to a certain extent, for the conflicts that have arisen around this method of classification. For it is from within an entirely genetic and historical linguistics that a general linguistics has developed in the last several decades. Because this general linguistics wishes today to free itself from the historical perspective and to make the synchronic study of languages prevail, it is sometimes led to take a position against the genetic principle of classification in favor of other methods. It would be of interest to ask to what degree these doctrinal differences affect the problem we are considering. A classification, no matter what it is, must begin by setting up its criteria. Those of genetic classification are historical in nature. One aims to explain the similarities-and also the differences-that have been observed among languages of a certain area, as well as others which are less apparent, by demonstrating their common origin. Proceeding from the data, the linguist employs a comparative and inductive method. If he can avail himself of ancient, intelligible, and rather extensive pieces of evidence, he endeavors to restore the continuity between successive states of a language or of a whole body of languages. From that continuity it can often be inferred that languages which are distinct today were derived from a single language. The proofs of this relationship consist in regular similarities, defined by equations between complete forms, morphemes, and phonemes. The equations in their turn are arranged in series, the more numerous the closer the relationship. In order for these equations to be conclusive, one must be able to demonstrate that they are due neither to chance coincidences nor to borrowings by one from the other of the languages considered or by these two languages from a common source, nor to the effect of convergences. The proofs will be decisive if they can be grouped into a bundle. Thus the correspondence between Lat. est: sunt, Germ. ist : sind, and Fr. e: so implies at one and the same time phonetic equivalences, the same morphological structure, the same gradation, the same classes of verbal forms, and the same meaning; and each one of these identical features can be subdivided into a certain number of features which are equally concordant, each one of which will evoke parallels in other forms of these

The Classification of Languages

languages. In short, there is here a conjunction of conditions so specific that the presumption of relationship is established. This method is well known and has been tested in the establishing of more than one family. There is proof that it can just as well be applied to languages without a history whose kinship is observed today, whatever their structure may be. A good example was given by Bloomfield in his comparison of the four principal languages of the Central Algonquian group, Fox, Ojibwa, Cree, and Menomini. On the basis of regular correspondences, he had shown that five different consonantic clusters \vith k as their second element had developed in these languages and had reconstructed in Primitive Central Algonquian the prototypes ck, sk, xk, hk, and nk. But one correspondence, limited to the form, 'he is red,' caused difficulty; it was represented in Fox by meskusiwa, in Ojibwa by miskuzi, in Cree by mihkusiw, and in Menomini by mehkon, with the Fox and Ojibwa sk making an anomalous equation to the Cree and Menomini hk. For this reason he had postulated a distinct rk group in Proto-Algonquian. It was only afterward that he chanced to study a Cree dialect of Manitoba in which the form in question appeared as mihtkrusiw with an -htk- group distinct from -hk-, thus justifying after the event the -ck- that he had assumed for theoretical reasons. 1 The regularity of phonetic correspondences and the possibility of predicting certain evolutions are not limited to anyone type of language nor to anyone region. Thus there is no reason for imagining that "exotic" or "primitive" languages will demand criteria of comparison different from those for the Indo-European or Semitic languages. The demonstration of primordial kinship implies an effort, often long and difficult, of identification applied to all levels of analysis: isolated phonemes, then bound ones, morphemes, complex signifiers, whole constructions. The process is connected to the consideration of the concrete substance of the elements compared: in order to justify the grouping together of Lat. fere- and Sans. bhara-, I must explain why Latin has precisely f just where Sanskrit has bh. No demonstration of kinship can escape this obligation, and a classification must collect a great number of these substantial points of sameness in order to assign each language to its place. Here again the conditions are valid everywhere and are necessary to the demonstration. We cannot, however, set up universal conditions as to the form a classification will take when applied to languages whose kinship can be proved. The image we have of a genetic family and the position we assign to languages grouped within such a family reflect, in reality-and it is well to be aware of this-a model of a particular classification, that of Indo-European languages. One will readily agree that it is the most complete, and for our purposes, the most satisfying. Linguists seek, consciously or not, to imitate this model each

88

PROBLEMS IN GENERAL LINGUISTICS

time they attempt to define the groupings of languages less well known, and it is so much the better if they are urged by this to show themselves increasingly rigorous. First, however, it is not certain that the criteria employed in Indo-European all have universal value. One of the strongest arguments for establishing the unity of Indo-European has been the similarity of the numerals, which remain recognizable today after more than twenty-five centuries. But the stability of these words depends perhaps on specific causes, such as the development of economic activity and exchanges (a development which occurred at a very early date in the Indo-European world), rather than on "natural" or universal reasons. In fact, it sometimes happens that the names of numbers are borrowings or even that the whole series of them is replaced, either for convenience or for some other reason. 2 Next, and most important, it is not certain that the model constructed for Indo-European is a constant type of genetic classification. \Vhat is peculiar to Indo-European is that each language participates to an approximately equal degree in the common language. Even when \ve take innovations into account, the distribution of the essential features of the overall structure is appreciably similar in languages of the same degree of antiquity, as has been confirmed in the case of Hittite and as can be assumed from the little that is known of languages like Phrygian or Gallic. Let us now see how the characteristics common to the languages of another well-established family, Bantu, are distributed. The Bantu area has been divided into geographic zones and each zone contains groups of languages that share certain phonetic and grammatical features; within these groups certain aggregates have been distinguished which have been subdivided into dialects. The classification is quite provisional, grounded on a very uneven documentation. Let us take it as it is, with some of the characteristics which distinguish those zones: 3 Northwest zone: monosyllabic prefixes; verbal inflection less developed than elsewhere; nominal prefixes with a particular form; North zone: dissyllabic nominal prefixes; a prefix-type of locative formation; a great abundance of augmentative prefixal formations; Congo zone: prefixes generally monosyllabic; vocalic harmony; development of verbal derivatives with an unusual compounding of suffixes; a generally complicated tonal system; Central zone: monosyllabic and dissyllabic prefixes; nominal classes for the augmentative, diminutive, and locative; a great development of verbal derivatives; a great development of ideophones; a system of three tones; East zone: a relatively simple phonetic system, a system of three tones, simplified verbal forms, a locative formation intermediate between prefixation and suffixation;

The Classification of Languages

Northeast zone: same characteristics with a more simplified morphology under the influence of Arabic; East Central zone: forms the transition between the Central and East zones; Southeast zone: monosyllabic an valere gives a delocutive function to valere in this expreSSIOn. We naturally think of the analogous situation of the Greek infinitive khairein. On the one hand we have the infinitive in its normal function: khairein tdll' ego s' ephiemai 'I give you leave to take your pleasure regarding all the rest';6 but khairein in its formulaic use in khairein tini 1~r;eill 'to send

Deloeutive Verbs

one's greetings to someone' represents the delocutive form that transposes the imperative khalre 'greetings l' The creation of delocutive verbs is effected under the pressure of lexical necessities and it is connected with the frequency and importance of pregnant formulae in certain types of culture. Latin offers several examples that are very instructive in their diversity. If negaTe is substantially derived from nee, it is insofar as it signifies 'to say nee' (with the emphasis on to say). Here again the basic term is one that forms an entire locution, in this case nee, conveying a negative judgment and constituting a sentence all by itself. Another delocutive is autumare, which is properly 'to say autem,' hence 'to argue, to assert.' It cannot be imagined that particles like nee or autem would have given rise to derived verbs if they had been taken in their logical function. It is only as formal elements of discourse that nee or autem lend themselves to the formation of verbs. Those verbs which have the exclusive connotation of "to say ... " are delocutives in the strictest sense. We know that Lat. quiritare 'to call for help' is to be explained literally as 'to shout Quirites!' For that we have the testimony of Varro: "quiritare dicitur is qui Quiritium fidem clamans implorat,'" and in addition, several examples of quiritatio, in the form of the appeal Quirites! or porro, Quirites!, have been preserved in literature. 8 Such a verb can only be delocutive since the base term is not the designation Quirites, but the call Quirites! Otherwise quiritare, if it were a denominative, would have to mean "to make someone a Quirite." The difference is obvious. We shall find in this mode of derivation the means for better understanding the meaning of an important term in ancient Roman ritual, the verb parentare 'to make a funeral oblation in memory of someone.' The relation with parens is obvious, but how should it be interpreted? A denominative parentare from parens would have to mean '*to treat someone as parens,' which omits the essential point. How then does it happen that the verb is restricted to funerary use? No one seems even to have seen the difficulty. It can be solved by an induction which we shall base on the following text. At the death of Romulus, or rather, after his sudden disappearance, Livy tells us that the people were at first seized with fright: deinde a paueis initio facto deum deo natum, regem parentemque urbis Romanae salvere universi Romulum jubent 'Then when a few men had taken the initiative, they all with one accord hailed Romulus as a god and a god's son, the King and Father of the Roman City.' A careful reading of this passage, in the midst of a narration that is so rich in authentic traditions, can uncover in Livy's formulation an expression which was certainly taken from an archaic ritual. With the expression parentem salvere jubent, it seems to us that we can restore a solemn formula that consisted of the appeal, 'parens, salve!' Livy preserves for us the

PROBLEMS IN GENERAL LINGUISTICS

very formula of the eonclamatio in the indirect syntax. This hypothesis becomes a certainty when the very same expression is found in a famous episode: during the celebration of the anniversary of the death of Anchises arranged by Aeneas, after the funeral games, when all the rites have been performed, he casts flowers upon the tomb of his father and utters the words, sfllve, sancte parens, iterum. 10 The agreement would appear decisive. This rite is exactly that of a parentatio. Here we find the explanation of parentare, which must signify literally, 'to utter the formula salve, parens!' The locution is reduced to its essential term, parens, on which parentare was formed, a typical delocutive.ll All that has been said of the relationship between Lat. salus and salutare is also valid for Fr. salut and saluer, as well as for the corresponding pairs in other Romance languages. It is the same relationship of locution to delocutive, and this relationship has to be posited synchronically without regard to the historical descendence from Lat. salutem to Fr. salut. It is not difficult at present to put into the same class Fr. merei and (re)mereier (O.F. mereier). We know from childhood that remereier means 'dire merci'; nevertheless it is important to stress the relationship through 'dire (and not faire) merci.' Merei in its lexical sense of 'grace' (cf. demander merei) would produce a denominative (re)mereier in the sense of 'to do a favor, to reprieve,' which is never the case. Only merei! as a conventional locution permits the justification of (re)mereier, which thereby is characterized as a delocutive, not as a denominative. Yet it must not be thought that the use of merci! as a locution was necessarily to lead to the creation of a verbal derivative like remereier. One could have recourse to distinct expressions. Such for example is the situation in Russian, in which the formula spasibo! 'thank you' has not produced a derived verb and remains independent of the verb blagodarit' 'to thank.' On the other hand, Eng. to thank and Germ. danken are clearly delocutives in relation to the substantives thank(s), Dank. Even in Gothic the locution pank fairhaitan (= *Dank verheissen), to translate Gr. khdrin ekhein (Luke 17: 9), shows that pank had become a fixed term, henceforth detached from pagkjan 'denken.' Since the basic term is taken so to speak as the name of the notion and not as the expression of the notion, modern languages retain the possibility, illustrated above by Lat. negare and autumare, of constructing a delocutive from a particle, on condition that this particle can be employed as a locution. Thus we have in English to hail 'to shout Clhail!" '; to encore 'to shout lIen_ core!" '; in American English, to okay and even to yes;12 in French, bisser 'to shout, IIbis!" , One might cite in Old High German a verb aberen 'to repeat,' taken from aber, like Lat. autum are from autem. We shall also consider Fr. tutoyer and vouvoyer as delocutives since they signify nothing other

Delocutive Verbs

243 than Ito say tu (vous).' It is obvious that a denominative of tu would be impossible; tu is not a quality that one can confer-it is a term of address of which tutoyer is the delocutive. Most of the verbs cited up to now are related by conventions to social life. Since the general conditions of culture are highly similar in the various modern western societies, it may seem natural that we encounter the same expressions in several languages. But the similarities that have been observed might be the result either of independent creations or, on the other hand, of influences of one language upon another. It would not be without interest to be able to state precisely the exact nature of the process in each case. Meanwhile the definition given here for delocutive verbs enables us to make the necessary distinctions. Thus in Gothic the adjective hails Ihealthy, in good health' has a formulaic use in the term hails Ikhaire! greetings!' But the derived verb hailj"an only means Ito cure'; it is a denominative. There is no hailj·an I*to greet.' It is in a more recent phase of Germanic that there appears a new verb, O.H.G. heilazzen, O.Ice. heilsa, O.E. halettan Ito hail,' which is a delocutive. It was probably created on the model of Lat. salutare. We notice also that Slavic agrees with Latin in the relation of O.Slav. cela (Russian celyi) Isalvus' to celovati 'salutare' (Russian celovat' 'to embrace'). Is this an independent creation in Slavic? The answer emerges from the very definition of the delocutive. The existence of an adjective celli is a necessary but not the only condition for the creation of a delocutive celovati; in addition it is necessary that the base form be susceptible of a formulaic use. Now we do indeed have in Slavic the equivalent of the Lat. salvus, but not that of Lat. salve! It is thus highly likely that the correlation celli: celovati was a Slavic calque from Latin, directly or through Germanic. The same question might be posed and solved in connection with a similar concordance between Armenian and Iranian. In Armenian there is druat leulogy, praise,' and druatem 'to greet, praise, acclaim,' like Lat. salus: salutare. Now this term comes from Iranian (Avest. druvatat- 'salus').13 From this one might jump to the conclusion that Armenian took the present derivative from Iranian as well as the noun. But it is to be observed that if Iranian has indeed converted the noun drud Ihealth' into a formula of salutationMiddle Persian drud abar to Igreetings to you!'-it has only druden- as a delocutive verb. It follows that druatem was created in Armenian itself through autonomous derivation. Ultimately, it is the resources and the structure of each linguistic system which decide this possibility of verbal as well as all other derivations. It is instructive to observe from this standpoint the differences of behavior ~ong languages starting with a common lexical situation. Let us take for

PROBLEMS IN GENERAL LINGUISTICS

example an expression with the same sense in three languages: Germ. willkommen, Eng. welcome, Fr. bienvenu. It is its use as a formula of welcoming which has determined its development in each language. The Germanic expression was so closely associated with a rite of welcome that it became, borrowed by O.F. wilecome, It. bellicone, the name for the large cup of hospitality. Now a delocutive is realized in English in the verb to welcome 'to say "welcome!" , German has not gone that far; there is no verb *willkommen but only a locution, willkommen (adj.) heissen 'to bid someone welcome.' In French the language has encountered a difficulty which it has only partially surmounted. There was a reluctance to derive from the adjective bienvenu, obvious in meaning and formerly separable (tres bien venus soies, 13th c.), a delocutive, *bienvenir (quelqu'un) , which would have been the exact equivalent of to welcome (someone). But a step has been made in that direction by the creation of an infinitive, bienvenir, which is limited to the expression se faire bienvenir de quelqu'un. The point of departure is the expression etre bienvenu (de quelqu'un), treated as a passive, on which a causative se faire bienvenir has been established in the same way that etre bien vu leads to se faire bien voir (de quelqu'un). But these are only approximations of a delocutive that was not achieved. Nothing is apparently more simple than the sense of the Latin benedicere 'to bless,' starting with the two morphemes that compose it, bene and dicere. This example has a particular interest for the present analysis because the form actually contains dicere and makes us suspect the condition of a delocutive. But examination reveals a history that is much more complex and less lineal, and whose description remains to be made. We shall limit ourselves for our purpose to indicating the most salient points. I. There was a use of bene dicere that has not been pointed out. It occurs in a passage in Plautus: quid si sors aliter quam voles evenerit?-Bene dice! 'what will happen if your lot turns out other than you wish?-No evil forebodings!' 14 Here Plautus with that locution bene dicere is certainly imitating Gr. euphimei! Moreover, nothing proves that this bene dice! ever led to a verb bene dicere in the sense of Gr. euphemein, because even in Greek there was no verb euphemein but only an infinitive euphemein, a transposition of the imperative euphimei (euphemeite) in an expression like euphemein keleuein 'to invite to utter words of good omen,' which is the ritual formulation of "to invite silence."15 2. The sense of the formula bene tibi dico 'I wish you well'16 is different. We must not think, as seems usually the case, that bene dicere literally signifies 'to wish well'; dicere here is not to be taken absolutely and, besides, has never signified 'to wish.' Bene must be understood as the objective term of dicere: "bene!" dicere alicui 'to say bene! to someone.' This bene! is an interjection

Delocutive Verbs

245

of a wish such as is met in many expressions: bene mihi, bene vobis 'to my good health, to yours!" in Plautus;17 bene nos; patriae, bene te, pater, optime Caesar 'to our good health, to yours! father of the country!' in Ovid,18 etc. From the very fact that the two components retain their autonomy, bene dicere did not succeed in taking the place of an authentic declocutive as would have been a verb derived directly from bene! We might imagine a German delocutive *pros(i)tieren which would give an idea of it. 19 3. A third acceptation appears when bene dicere is used in the classical language for 'to praise, to eulogize someone.' This is again a development that is due to a literary influence: bene dicere serves to translate Gr. eulogein, which is quite different from euphemezn. 4. Finally, when Gr. eulogezn was itself chosen to render Hebr. brk, it was benedicere (now becomes a single sign) which remained the Latin equivalent for it, but this time in the new Judaeo-Christian value of 'to bless,' producing in its turn benedictus and benedictio. This is the modern notion. In order to complete the characterization of this type of verbal derivation, it seems convenient to ward off two possible confusions. In the first place, we must distinguish carefully between delocutives and verbs derived from interjections: Fr. claquer, huer, chuchoter, Engl. to hush, to boo, etc. A delocutive always has as its root a signifier that can be interjected into discourse without ceasing to be a signifier, while verbs like claquer are constructed on simple onomatopoeias. Here the distinction is easy. A little more insidious is the temptation to confuse the delocutives with what are called "verbs of wishing" in the traditional grammars. Of course expressions like welcome! and salut! serve to transmit a wish. But this psychological background is alien to the problem. The delocutive is defined not by the intentional content but by the formal relationship between a locution and a verb denoting the utterance of that locution. The meaning of the constituent locution matters little. The difference appears clearly if we compare the "verb of wishing" par excellence, to wish, with a delocutive like to greet. The word wish is not a formula of wishing; it is a substantive like any other, and the derived verb, to wish, is a simple denominative, while greetings is certainly a substantive, but also in the form greetings! a formula of greeting; that is why to greet, signifying 'to say "greetings!" , will be called a delocutive. One should also classify as delocutives Fr. sacrer 'to say, "sacre! ... " " pester 'to say, ''peste!'' , The essential and signal feature of a delocutive is that it is in the relationship of "to say . . ." with its nominal base, and not in the relationship of "to do ... ," which belongs to the denominative. It is not the least instructive characteristic of this class to show us a sign of language deriving from a locution of discourse and not from another sign of language; by this very fact, delocutives are, above all, from the moment at which they are created,

PROBLEMS IN GENERAL LINGUlS~

verbs denoting activities of discourse. Their structure as well as the rea~ that summon them into existence assign them a very particular position am the other classes of verbal derivatives.

From Studia Philologica et Litteraria in Honorem L. Spitzer, ed. A. G. Hatcher K. L. Selig, Bern, 1958, pp. 57-63

Lexicon and Culture

TWENTY -FOUR

Semantic Prol)lems in Reconstruction

THE mrAS OF SE:'IL\~TI(,S ;HC ~till presented in such vague form that in order to treat any aspect of the fleld, we should begin by establishing a body of rigorous definitions. But these deflnitions would in their turn require a discussion of the yery principles of meaning. This would be a long and arduous task and the works devoted to semantics up until now give only a faint notion of it. And so in this article, v,·hich is limited to the theme suggested by the editors of this journal, we shall have to proceed empirically, on the wholc, neglecting for the momcnt theoretical considerations in order to treat concretely several typcs of problems which the linguist encounters when he concerns himself with reconstruction. In general, the criteria of a formal reconstruction can be strict because they stem from precise rules that cannot be set aside unless one is in a position to substitute more exact rules for them. The whole apparatus of phonetics and morphology enters in to sustain or refute these endeavors. But when it is a matter of meaning, one has as a guide only a certain probability based on "common sense," on the personal evaluation of the linguist, and on the parallels that he can cite. The problem is always, at all levels of the analysis, within just one language or at different stages of a comparative reconstruction, to determine if and how two morphemes which are formally identical or similar can be shovm to coincide in meaning. The only principle we shall use in the considerations that follow, by taking it for granted, is that the "meaning" of a linguistic form is defined by the totality of its uses, by their distribution, and by the types of associations therefrom. In the presence of identical morphemes with different meanings, one must ask oneself whether there is some use in \vhich the two meanings converge. The answer is never given in advance. It can only be found after a careful study of all the contexts in which the form may appear. One does not have the right to assume what the answer is, either positively or negatively' on the basis of likelihood. I. Take, for example, the case of the English homophones story 'narrative'

PROBLEMS IN GENERAL LINGUISTICS

and story 'a set of rooms.' What keeps us from saying that they are the same is not our feeling that a "narration" and a "floor" are irreconcilable, but the impossibility of finding any usage of such a nature that the one meaning might be interchanged with the other. Even expressions purposely chosen for their ambiguity like to build a story or the third story (of an anthology, of an apartment building), immediately lose their ambiguity, once they are placed in an authentic context. We must then consider them distinct. It is only by way of confirmation that the etymological proof will be utilized: story 'narrative' < O.F. estoire (historia), but story 'floor' < O.F. estoree (*staurata). We may not have the etymology, but even if we did, it would not by itself suffice to guarantee the present independence of two morphemes, which could, by virtue of their identical forms, have become associated through their meaning in some way and have created a new semantic unit. 2. Here is the opposite case. In French we have voler 'fly' and voler 'steal.' The two verbs are distinct in every respect. Voler 'fly' belongs to the semantic class of "walk, run, swim, creep," etc.; voler 'steal' is synonymous with "rob, purloin," etc. Voler 'fly' is intransitive; voler 'steal' is transitive. Derived forms include only one term which is common to both verbs: vol. Otherwise they differ: voler 'fly' produces voleter, s'envoler, survoler, volee, volatile, volaille, voliere; but voler 'steal,' only voleur. This very limitation of voler 'steal' makes us suspect that it goes back to a specialized use of voler 'fly'. The condition for it would be a context in which voler 'fly' would lend itself to a transitive construction. This context is found in the language of falconry; it is the expression, "Ie faucon vole Ia perdrix" (= reaches and seizes in flight). Such is the actual condition, not presumable in advance, in which the exceptional transitive use creates a new sense for voler; in this situation the vol of the bird means both "flying" and "stealing." The coexistence of the two volers should not lead us to reconcile them in an improbable unity; the peculiar situation of one of these two homonyms and especially the scarcity of its derived forms urge us to seek the typical usage that divided a unitary semantic field into two distinct parts. 3. In the evaluation of the differences in meaning that intervene among the members of a formally bound ensemble, the linguist is always inclined to let himself be guided unconsciously by the categories of his own language. Hence the semantic problems that, all things considered, come down to problems of translation. We encounter this even in those reconstructions that have never been challenged and that might pass for obvious facts. The correspondence between Gr. tithemi, etheka 'to put down' and Lat. facere 'to make' is an elementary datum of comparative linguistics. From that we might conclude that *dhe- admits both the sense of "set" and that of "make." But the connection between "set" and "make" is not so clear to us that

Semantic Problems in Reconstruction

without proof we may accept it as true for Indo-European. In our classification of ideas, "set" goes with "place, put, lodge," etc.; and "make" with "accomplish, construct, manufacture, operatt!," etc. The two lines do not meet. The very multiplicity of the acceptations for "make" does not seem to help us clarify the connection that is nevertheless involved in these ancient associations. In order to justify this sense-relation, linguists have had recourse to technical uses. l Actually the reasons must be sought in a more precise definition of the uses. First, it should be observed that even in a case in which the translation "set" is admissible, the conditions of use show that "set" properly signifies "set down something which will last from now on, which is destined to endure": in Greek, with themeiZia 'to set the foundations, with bomon 'to raise an altar.' That is why it is suited to signify lito establish in existence, to create"; cf. in Old Persian biimim ada . .. asmanam ada 'he has set (= created) the earth, he has set (= created) the sky,' in Greek khdrmat' etheken 'he has set (= created) joys for men' (Pindar OZ. 2. 101), etc. In the second place, it will be noticed that one of the most frequent constructions of *dhe- is predicative, which furnishes the usual sense of "make," both in those languages which also have "set" and in those, like Latin, which have only "make": basilea tina thelnai is literally aliquem regem facere, and an expression like theinai tina athdnaton is the exact equivalent of immortalem facere. It is enough to indicate the principle; the examples are abundant. The important thing is to see that: (I) the distinction between "set" and "make" does not correspond to the Indo-European reality in the settled form it has for us; (2) the syntactic construction of *dhe- is an essential component of the use and the meaning; (3) the notion of "make," insofar as it is expressed by *dhe-, is determined by particular connections which alone allow for defining it, for the definition is possible only in the terms of the language itself. 4. This situation often occurs in less recognizable forms. This presents difficulties that may arise from the fact that one or the other of the meanings considered is inexactly or too cursorily defined. We shall take as an example of this a Greek verb whose meanings do not seem to have created any problem until now. There is in Greek the verb trepho 'nourish,' with numerous derivatives and compounds attesting to the same meaning: trophds 'fosterfather,' tropheus 'stock-raiser,' trophe 'nourishment,' dio-trephis 'foster-child of Zeus,' etc. It is said to be identical with trepho 'to thicken, to congeal (a liquid),' perf. tetrophe 'to coagulate, to be compact,' which in its turn has been attached to thrdmbos 'blood-clot' (in spite of the phonology), then to a series of unrelated comparisons which one will find detailed in Boisacq 353 and which will not delay us here. The only important thing for us is the relationship in Greek itself of trepho 'nourish' to trepho 'curdle (milk).' It is

PROBLEMS IN GENERAL LINGUISTICS

actually quite possible that the two meanings are only one, but how to check it? The dictionaries do not indicate any trouble. That of Liddell-Scott-Jones defines trepho thus: "I. thicken or congeal a liquid; 2. usu. cause to grow or increase, bring up, rear, esp. of children bred and brought up in a house." The same in Bailly: "I. rendre compact; 2. rendre gras, engraisser, nourrir." Even for someone who trusts only in the "feeling" of a language, such a relationship must seem so strange as to demand a verification of the uses. That it could have been thought obvious that "to curdle (milk)" leads to the sense of "to nourish, to raise (a child)" is enough to discredit that "intuitive" empiricism which serves as the method for most reconstructions. Here the disparity of meanings seems such that a reconciliation of them could be nothing more than contrived. In reality, the translation of trepho by 'nourish' in the use that is actually the most common does not suit all the examples and is itself only an acceptation of both a broader and a more precise sense. In order to account for the ensemble of semantic connections of tripho, we have to define it as: 'to encourage (by appropriate measures) the development of that which is subject to growth.' With paldas, hippous, we would translate it as 'to nourish, to raise (children, horses).' But there is also triphein aloiphen 'to encourage the increase of fat' (Od. 13. 410); trephein khaiten 'to let his hair grow' (Il. 23. 142). It is here that a peculiar and "technical" development is inserted, and it is precisely the sense of "curdle." The Greek expression is triph£in gala (Od. 9. 246), which must now be literally interpreted as 'to encourage the natural growth of milk, to let it attain the state towards which it is tending,' or, prosaically, 'to let it curdle.' This is nothing other than an idiomatic connection of triphein to the sense of 'to let grow, to encourage growth' that it has everywhere. From the standpoint of Greek, there is no difference between trephein khaiten 'let the hair develop itself' and trephein gala 'let the milk develop.' Nor is there any difference between trophies paldes 'children who have grown (and attained the adult age), and kumata trophoenta, kuma trophi 'waves that have attained their full development.' There is thus no longer any problem as to the classification of the two meanings of -trepho since there is only one meaning, which is the same everywhere. One can conclude that trepho 'to clot,' does not exist; what does exist is a use of tripho gala that creates an association that is unusual to us but explicable in the Greek contexts. It can also be seen that the whole difficulty comes, basically, from differences between the lexical resources of the languages being considered. While triphein palda translates directly into English or French (,rear a child, nourrir un enfant'), trephein gala requires a specific translation ('curdle milk, cailler du lait'). The linguist who asks himself, "How to reconcile 'curdle' and 'rear' or 'cailler' and 'nourrir'?" or who invents an affiliation between these two meanings, is the victim of a false prob-

Semantic Problems in Reconstruction

253 lem. The question does not arise in either a modern language, in which the forms are different, or in Greek, in which the meanings are identical. This is only one among the many gratuitous difficulties created in semantic reconstruction either by an insufficient definition of the terms under discussion or by an unjustified transposition of the values from one semantic system to another. 5. The same problem could be posed, no longer within a historical language, but in the synchrony of a formal reconstruction. In Indo-European there is a root *dwei- 'to fear' well attested by Greek deos 'fear' (*dweyos) and the perfect de-dwoi-a 'I have fear' furnishing the present deido, by Av. dvae()ii'menace, cause of fear', by modern Arm. erknl'im 'I fear.' This *dwei- 'fear' is substantially the same as the stem of the numeral *dwei- 'two.' The resemblance persists in derivatives of historical times: Hom. de-dwoi-a 'I have fear' appears to have been built upon the same stem as the adjective dwoi-ds 'double,' and Arm. erknc'im 'I fear' recalls erku 'two' (*dwo); the gradation in the Homeric perfect 1st sing. de-dwoi-a : 1st pI. de-dwi-men is in conformity with that of the numeral *dwei- (*dwoi-) : *dwi-. In short, everything seems to suggest that the forms of these two roots are identical. Is this chance? But in order to exclude chance, it must be demonstrated that the establishment of identical forms is verified in the meaning. And, what connection could one imagine between the meanings of "fear" and "two" which would not look like mere cleverness? It is nevertheless necessary to consider this more carefully, and not reject without an examination the possibility of a relation. For-this is essential-if we can regard the notion of "two" as "simple," we have no right to assume that a notion like "to fear" is equally "simple." Nothing gives us a priori assurance that it had the same semantic structure in the ancient stages of Indo-European as in the language of the present discussion. And the analysis of this semantic structure itself has as a condition the study of the uses of *dwei- 'to fear' in situations in which we can best observe them. Homeric Greek lends itself rewardingly to such a study. For it is in a text of the Iliad, notwithstanding that it has been read and reread a thousand times, that a solution that had not been offered till now presents itself. Here is the passage: lien mega pema . . . eisordontes deidimen; en doiei de saosemen e apolesthai neas (Il. 9. 229-230), literally: 'foreseeing a great disaster, we are afraid (deidemen); what is in doubt (en doiei) is: shall we save or shall we lose the ships?' The text itself, bringing deidemen and en doiei together in the same sentence, illuminates their relation as if by a classroom demonstration. The expression en d(w)oyei 2 (esti) properly signifies 'the thing is in double, in doubt, in dubio,' that is, 'it is redoubtable.' Hence it follows that *dwei- 'to fear' signifies 'to be of two minds [double], to doubt'; cf. O.F. douter, Mod. Fr. redouter 'to fear.' The situation

254

PROBLEMS IN GENERAL LINGUISTICS

described in the text (the feeling when faced with a perilous alternative) restores the connection being sought for between *dwei- the numeral and *dweithe verbal. Henceforth they can be considered as identical in meaning. As further proof we shall use parallels like Lat. duo, dubius (in dubio esse), dubitare; Germ. zwei, zweifeln, etc. Thus, thanks to a decisive context, a notion like "to fear" in Indo-European takes shape with its specific connections, which only its use can reveal and which are different from those that determine it today3. 6. The necessity of making use of contexts would seem to be a methodological principle too obvious to deserve insisting upon. But when we reduce meaning to varieties of use, it becomes imperative to ascertain that the uses not only allow apparently different meanings to be brought together, but also allow us to state the reason for their difference. Factors that provoke the emergence of a new "species" of meaning should also enter into the reconstruction of a semantic process. Lacking that, the perspective is falsified by imaginary evaluations. We shall take an example from the most commonplace comparison of them all, that of Lat. testa and Fr. tete. Handbooks go on repeating that the passage of the meaning of testa 'crock, piece of pottery' to that of tete was due to jest. The explanation is found even in the most recent dictionaries. 4 It is time to look at the facts, which, moreover, are obvious; only they have not been considered. The problem begins with the term for "head" in classical Latin. It has been observed that caput signifies not only 'head' but also 'person' and 'capital (financial), and 'capital (of a province),; it enters into liaisons such as caput amnis 'source (or mouth) of a river,' caput coniurationis 'head of the conspiracy,' caput cenae 'principal dish of a meal,' caput libri 'chapter in a book,' caput est ut ... 'it is essential that ... ,' etc. The number and extent of these variants weakened the specific meaning of caput 'head,' which led to two possible solutions. Either it could have been redetermined as *caput corporis, which should itself have been ambiguous and which, in any case, the language rejected, or it could have been replaced by a different term. This is what happened in Latin: the language resorted to testa, a term that used to designate any hard shell and which was first applied to what we call today the "brain-pan" (cf. Fr. boite cranienne, Germ. Hirnschale). The sense of "skull" appears clearly in late LatinS (Antoninus Placentius: vidi testam de homine 'I saw a man's skull') and it was used that early as a name for the "head"; 'testa: caput or vas fictile' (C.G.L. 5 :526-539), whence Old French teste 'skull.' It is probable that, as an anatomical term, testa was in use among Roman doctors long before the texts mention it. And so there is in this development neither a joke nor, as a matter of fact, a peculiarity worth noting. One might even find that the case of testa: tete has usurped its place in the traditional teaching; it simply offers another special aspect of the refurbishing that reached most of

Semantic Problems in Reconstruction

255

the names for the parts of the body. From this we can trace the successive oppositions: Lat. caput: testa> O.F. chef: teste> Mod. Fr. tete: crane. But in this corrected perspective, the considerations of testa as a humorous designation no longer seem to have any foundation. The real question would rather be to study how caput and testa coexisted and were delimited in late Latin, and chef and teste in Old French, in order to arrive at their current distribution. If this research remains still to be done, it is at least partly because an inexact evaluation of the nature of the process has obscured its importance. 7. Within the framework of a grand-scale comparison of several languages, one often observes that forms that are obviously related are each distinguished by a particular variety of meaning. Although the semantic unity of the family is undeniable, it does not seem capable of being defined exactly. One has the impression that the "primary meaning," preserved exactly by one language, deviates for particular reasons in each of the other languages, thus producing a composite image of the semantic situation. In general the comparatists do not take the time to examine this image when the formal correspondences are satisfying. Or if they do consider the development of one of the forms, they do so without regard to the question of the ensemble. Such, for instance, is the case of the term for 'road' : Sans. pdnthiil}, Av. pantd, Arm. hun, O.Slav. PQi'i, O.Pr. pintis, Gr. pantos, Lat. pons. The Indo-European antiquity of the terms is guaranteed by the archaisms in the inflection. We cannot say that the meaning raises an obstacle to the reconstruction of a common form. Nevertheless the divergences appear serious enough to justify an examination. The Indo-Iranian, Slavic, and Baltic words mean 'road.' But Gr. pantos signifies 'sea'; Lat. pons designates 'bridge,' and Arm. hun, 'ford.' Since these meanings are not equivalent and since, in the dialectal distribution, it is especially in Greek and in Latin that the divergence shows itself, one would tend to think that this disagreement results from reasons of style or of culture. In Greek one might think that poetic imagination assimilated the "sea" to a "road." In Latin, the transfer of "road" to "bridge" might be thought a result of a terra-marine civilization. These hypotheses stem from another hypothesis, not recognized as such, unformulated and unconscious: the hypothesis that the primary sense is that of "road" because it is documented in an ancient dialect like Indo-Iranian or because of the agreement between Indo-Iranian, Slavic, and Baltic, or by virtue of its "simplicity"; and the meanings of "sea" or "bridge" or "ford" would be deviations. But the uses we find in the most abundant ancient texts, in Vedic, 6 allow us to grant a more exact notion to pdnthiil} and to show its various shades of meaning. First of all, there are in Vedic several other terms for road, and they are all in some way distinguished from this one: yiina- denominates the 'road' of souls towards their rest (devayiina,pitryiina); miirga-, the path of wild animals (mrga);

PROBLEMS IN GENERAL LINGUISTICS

adhvan, the beaten track; rathya, the wagon track. What characterizes the pdntha/.z is that it is not simply the road as a space to cover from one point to another. It implies difficulty, uncertainty, and danger, it has unforeseen detours, it can vary depending on who is traversing it, and moreover, it is not just terrestrial-birds have their road, and rivers too. The pdntha/.z is thus neither plotted in advance nor regularly trod. It is indeed rather a "crossing" attempted over an unknown and often hostile region, a path opened by the gods to the onrush of waters, a passage past natural obstacles, or the route that birds invent in space; in short, a way into a region forbidden to normal passage, a means of going through a perilous or uneven expanse. The closest equivalent would be ucrossing" rather than uroad," and it is indeed this sense which explains the diversity of the documented variants. Starting with Sans. pathya and into the history of Indo-Aryan, we have 'road,' but this sense is no more "primordial" than the others; it is only one of the realizations of the general signification defined here. Elsewhere these realizations are represented differently. In Greek, the "crossing" is that of an arm of the sea (cf. Hellds-pontos), then more broadly, of an expanse of water serving as a "passage" between two continents; in Armenian, it is a uford"; and in Latin, pons will designate the 'crossing' of a stream of water or a dip in the ground, hence a 'bridge.' We are not in a position to give the exact reasons, which arise from geography or culture, for these particular detenninations, all of them prehistoric. But at least we can realize that "road," "arm of the sea," "ford," and "bridge" are variants of a meaning whose reconstruction is made possible by these terms, and that the problem does not refer to the semantic aspect of the term in one language or the other but concerns each one of them and the entire family of which they are members. 8. When in the comparison of the terms of a unitary group we find ourselves in the presence of developments that are distributed in clearly marked groups, we are often obliged to indicate in which direction the meaning has varied and which of the observed meanings has produced the other. It is then necessary to refer to a criterion that is general and fixed enough not to need to be proved each time. One of the most usual criteria is the "concrete" or "abstract" character of the meaning, the evolution being supposed to have proceeded from the "concrete" to the "abstract." We shall not emphasize the ambiguity of these terms, which have been inherited from an obsolete philosophy. The question is simply to know whether, even when accepted without argument, they can furnish a principle that would be valuable in semantic reconstruction. The best way to test them would be to examine the application that has been made of them-unconsciously-in a rather important lexical problem. This is the curious case of an etymological family, well defined in its formal relations, whose meaning is distributed among

Semantic Problems in Reconstruction

257 notions that are very material, on the one hand, and moral and institutional on the other. It is the term that, in general, has to do with "trust" and which in the Germanic Middle Ages had great social and cultural importance (d. trust, true, truce, etc.) The unity of meaning in the Germanic forms stands out even as they are enumerated. In Gothic we find trauan 'pepoithenai, to be confident,' ga-trauan 'pisteuesthai, to trust,' trauains 'pepoith~sis, confidence,' traustei (from the gen. trausteis) 'diatheke, pact, alliance'; in addition, O.lce. triia, O.E. triian, O.H.G. tril(w)en 'to have confidence,' derived from *triiwa, in O.Ice. trii 'respect,' O.E. truwa 'religious respect, belief,' O.Ice. triir 'faithful'; to the full development, O.E. treowian, O.H.G. triuwen 'to trust'; a derivative, *drou-sto-, gives O.Ice. traustr 'trusty, strong,' and the abstract *draustya, in Goth. trausti, O.Ice. traust 'confidence,' O.H.G. trost 'act of giving confidence, encouragement'; an adjective *dreuwo- in Goth. triggws, O.Ice. tryggr, O.H.G. gi-triuwi 'faithful,' and in the O.E. noun, treaw f., O.H.G. triuwa 'fidelity.' But outside of Germanic, the related terms convey an entirely different meaning, which is, however, partially represented in Germanic as well. They designate "tree," sometimes, in particular, the "oak tree," or "wood" in general: Gr. drus 'oak,' Sans. daru, dru-, Av. dru- 'tree, wood,' drvaeni- 'wooden,' Goth., triu 'wood, tree' (and the corresponding forms, Eng. tree, etc.), Gall. derw·pl. 'oaks,' O.Slav. drevo, Russ. derevo 'tree,' Lith derva 'pine wood.' How should this distribution of meaning, "tree;' on the one hand, "fidelity" on the other, be organized within a set of forms which are clearly related in other ways? This whole etymological family was studied by H. Osthoff in an important chapter of his Etymologica Parerga (1901) significantly entitled "Eiche und Treue." He establishes the Indo-European word represented by Gr. drus 'oak' as the point of departure of the whole morphological and semantic development, and claims that the moral values implied in Treue and truste proceed from it. The Gothic adjective triggws, O.H.G. gitriuwi 'getreu, faithful,' would, then, properly signify 'firm like an oak.' In the Germanic mentality, the "oak" would have been the symbol of firmness and confidence, and the image of "oak" would have inspired the whole set of the representations of "fidelity." For more than half a century Osthoff's theory has been considered established; etymological dictionaries refer to it as to a proven demonstration. 7 One would accordingly imagine this to be an illustration of a concrete designation evolving into a moral notion: an institution having a vegetable symbol as its origin. But as soon as this construction is examined, its deficiencies are revealed. Osthoff, by putting the name of the "oak" at the point of departure for the whole derivation, admits implicitly-the argument is essential for this theory

PROBLEMS IN GENERAL LINGUISTICS

-that the name for "oak" is Indo-European. But everything contradicts this. It is only in Greek that drii- signifies 'oak.' Everywhere else the sense is "tree, wood" in general: Bitt. taru, Ind.-Ir. dtiru-, drii-, Goth. triu, etc., O.Slav. druva pI. In Greek itself, d6ru applies to a tree (Od. 6. 167), to the wood of a ship (Il. 15. 410), to the wood of a lance, and to the lance. Even more to the point, the sense of "oak," which Gr. drus has in the classical language is secondary and relatively recent: a scholiast (ad Il. II. 86) still knew that "the ancients called any kind of tree drus (drun ekaloun hoi palaiol ... pan dendron). The generic term for "tree" denoted the most important tree, the "oak," probably under the influence of the beliefs attached to the prophetic oaks of Dodona. Moreover the common name for the tree, Gr. dendrewon, can be explained as a broken redoubling with dissimilation, from *der-drew-on (cf. Lat. cancer from *kar-kro), and rests on *drew- in the sense of 'tree.' Thus everything confirms that *dreu- designates tree in general and that the sense of 'oak' was acquired only in Greek. There is a reason for this limitation = the oak grows only in one part of the Indo-European area, in the middle part of Europe which goes from Gaul to northern Greece and not beyond there toward the east; in fact there is no Indo-Iranian term for "oak." Thus Osthoff's demonstration is weakened in its very principle: the signification that he believed to be the original one appears to be late and limited. Consequently the relation between the notions he set up loses its principal support. This must be carried further and the methodological flaw of the whole line of reasoning exposed. The morphological correlations and the distribution of the forms do not indicate a relation between the terms that denote "tree" and those for "fidelity" such that the second would derive from the first. They are equally distributed in each language and both depend upon the same signification, which allows itself to be reconstructed with the aid of the ensemble of attested forms. The formal base must be set up as I *der-wII *dr-eu-, with the sense of 'to be firm, solid, sound.' Cf. Sans. dhruva(for *druva- contaminated by dhar-), Av. drva, O.Pers. duruva- 'firm, sound,' Gr. dro(w)6n' iskhur6n (Hes.), O.Slav. *su-dorwa > sudravu, Russ. zd6r6v 'sound,' Irish derb (*derwo-) 'sure,' O.Pr. druwis 'faith' « 'security'), Lith. drutas 'firm, powerful,' etc. Here is the natural place for the Germanic members of this group like Goth. trauan, trausti, etc., which are directly derived from this base and which fixed the terminology of "confidence" in Germanic. Hence it is this common signification that the designation of "tree" shares in the same degree. Contrary to Osthoff's reasoning, we consider that *derwo-, *drwo-, *dreu- in the sense of 'tree' is only a particular use of the general sense of "firm, solid." It is not the "primitive" name of the oak which created the notion of solidity; on the contrary, it is by the expression of solidity that trees in general and the oak in particular were designated: Gr. drUs (Gall.

Semantic Problems in Reconstruction

259 derwen) signifies literally 'the solid, the firm.' There is a parallel in Iranian in which 'tree' is called draxt (M. Pers.), diraxt (Mod. Pers.), which goes back to Av. draxta-, the adjective of drang- 'to hold firm.' The romantic conception of the oak as the inspiration of fidelity gives way to a less peculiar and probably more exact represeI!tation: the name *dru- for tree has nothing "primitive" about it; it is a qualification which once attached to its object became its designation and was separated from its semantic family; hence the coexistence of two morphemes, which have become distinct, like tree and true in English. Here it can be seen how fallacious the criterion of the "concrete" and the "abstract" is when applied to a reconstruction, and how important is the necessary distinction between signification and designation. 9. The difference of meaning and the difficulty of reconstruction reach an even higher degree when the forms are distributed among distinct and grammatically irreconcilable classes. In the cases considered up to this point, we had to do with forms whose status, at least, was not at odds with a direct comparison, the meaning alone giving rise to discussion. But how to operate when the formal similarities are contradicted by functional differences? We can easily establish a relation between nominal and verbal forms that are distributed according to the principles of derivation. Can we bring together into the same semantic family forms of which some are particles and the others are verbals or nominals without any common syntactic use? Such a problem is posed, however, by the coexistence of forms of different series that group themselves around the Indo-European term *pot(i)-, which designates the 'chief.' In trying to solve it, we shall answer the question of method brought up by this case. An I.E. *pot(i)- is represented in a free state in Sans. pati- 'chief' and also 'husband,' Gr. posis 'husband,' compounded in Sans. jas-pati- 'master of a line of descendants' (a very productive Indo-Iranian type), Gr. des-potes, Lat. hospes, compos, Lith. viespats 'lord,' Goth. brup-faps 'bridegroom,' etc. Lat. potis and an ensemble of derivatives, potior, possum, possideo, can easily be connected to this. The meaning, uniformly distributed, is defined as 'master, chief,' with a development in Latin and Italic toward the notion of "power." But there is a homophony between this *pet-/*pot(i)- 'chief' and a particle of identity, *pet-jpot(i)-, which means 'same, self': Hitt. -pet, Av. -paiti, Lat. -pte, Lith. -pat. The two forms do not always occur together; Hittite does not have a form of *pot(i)- 'chief,' and the particle seems to be lacking in Sanskrit and Greek. But in the majority of languages both appear without there always being a discernible connection between them. The reconstruction of a semantic relationship must necessarily start with a decision of principle: which of these two classes should be taken as the point of departure? The question has been settled in opposite ways. Meillet decided

260

PROBLEMS IN GENERAL LINGUISTICS

that one should proceed from *poti- 'chief' and that the value of Lith. pats '(him)self' resulted from an appositional use, which Meillet did not go on to explain;8 this hypothesis is hardly compatible with the obvious antiquity of the particle. More likely, but not exempt from difficulties, is the opinion of H. Pedersen, who takes the meaning of "master" from "self," citing not exact proofs but parallels: he compares certain uses of "himself" to designate the "master of the house," such as Gr. autos, Lat. ipse, Dan. dial. han selv 'the mistress of the house,' Russ. sam, sama 'barin i barynja.'9 But all that these examples can prove is that in a very special situation, as is that of familiars or domestics, a pronoun suffices to refer to the personage of authority. Slaves in Greek or Latin comedy express themselves thus, but not free men in the solemn language of religion or poetry. The use of ipse for the master of the house is a simple fact of "parole," and it has never reached the level of "langue." Besides, it is too sporadic and recent to account for the obviously archaic and "noble" forms such as the Sanskrit pair pati/patni, and the Greek, posis/potnia. It has not been observed that this "ancillary" use of autos, ipse, etc., has ever produced a lexical denomination for "master" as such, or a derivation starting with this sense. In short, these parallels are both too limited in their sphere and too "familiar" in style for them to be seen as other than "situational variants." The pronouns ipse and autos can occasionally designate the master; they have never signified "master" outside their context. They do not assist us in recovering the connection between the two forms of *pot(i)-. The manner in which the forms of each series are distributed respectively is worthy of observation. It will be noted that Hittite, an archaic dialect in many respects, possesses only the particle -pet 'self' (apas-pet 'himself, he exactly'), and there is no trace of a nominal form like *pot(i)-. This makes one assume that *pot(i)- has a chance of being secondary. On the other hand, the nominal forms of the "master" group are not linked to any verbal root; when there is a verbal form, such as Sans.patyate, Lat. potior, it is clearly denominative. This is, therefore, a lexical family which is entirely and exclusively nominal. The terms under consideration are thus on the one hand a particle, on the other, a nominal form. I t is necessary first to state precisely the function of the particle -pet. There are two distinct expressions of equivalence in the Indo-European languages, which can be illustrated by the example of Gothic, which possesses both sama. and silba; by sama 'same' the equivalence is announced as a permanent fact of the object recognized in different aspects or in different instances; by silba 'self' the equivalence is opposed to otherness: "himself" to the exclusion of all others. It can be mentioned in passing that the value of emphasis and of contrast inherent in the expression of the "self" category

Semantic Problems in Reconstruction

26r

leads to its being signalled either by reference to the corporal being (hence, Ind.-Ir. tanii-, Hitt. tuekka-, O.H.G. leip, Fr. en personne, en chair et en os, etc.), or by an emphatic denomination like the superlative (hence Germ. selbst, Gr. autotatos, Lat. ipsissimus [ef. met-ipsimus > O.F. medisme, Fr. mente], Slav. sam as the superlative, etc.) as "exemplary" personifications of the notion. The function of the enclitic -pat in Hittite, -pat in Lithuanian, whose use is inherited, obviously answers to the notion of "self": Hitt. apas-pat Ithat one exactly, himself,' Lith. ten-pat Ithe self-same place,' as pats Imyself,' with a superlative value developed in Lithuanian: pats pirmasis Ithe foremost.' In this function, the particle is attached to the pronoun, and then there occurs a selective association, which appears clearly in Iranian, in which -pati is incorporated into the reflexive-Avo xvae-paiti- loneself,' and especially the derived xvaepai()ya-, O.Pers. (h)uviiipasiya lone's very own,' in the O.Pers. predicative construction (h)uviiipasiyam kar- Iproprium facere, to appropriate' -for any person whatsoever, but always for one person. From this use can be deduced the explanation of the nominal *pet/pot, suffixed and nominalized by -i in *poti, which means the person in his own right, the lIipse" with some sort of determination. Actually the derivative of pati-, Sans. patya-, constructed with the dative, preserves the sense of u to be proper to": asutis carur mddiiya patyate, lit. Ian agreeable beverage is proper to drunkenness' (R. V. 8. I. 26), and Av. pai()ya- signifies Ito have in one's own right' (and not Ito be master of). This definition of *poti as Ithe ipse, the being in person' is conditioned by the determination that actually always accompanies the term in the most ancient expressions: the *dems poti (Av. dmg pati-, Ved. dam-pati, Gr. des-potes) is literally Ithe ipse of the house, the being itself of the family,' the one who personifies the social cell. That is what we transpose into the terms of our own culture by the usual translation Imaster of the house.' From this proceed all the other compounds, which gradually come to include more territory: Sans. vis-pati, Av. vis-paiti-, Lith. vies-pats Ithe one who is the ipse of the *wik- = master of the clan,' etc. We find two indications that corroborate this interpretation. The sense of Lat. hospes (*ghos-pet-), which designates the one who receives as well as the one who offers hospitality, is better explained as the uipse" rather than the Ilmaster" of the reciprocal benefit designated by *ghos(ti)-, in which the two members are equal partners. In addition, it now becomes possible to connect the series of compounds in *-poti to a formation of the same meaning, but of different structure, which belongs to western Indo-European. Saussure once called attention to the curious formation of the following terms: Lat. dominus tribunus, Goth. piudans 'king,' kindins 'hegemon,' O.Ice. drottenn 'prince', they are all secondary derivatives in *-no- of base terms for designating chiefs:

260

PROBLEMS IN GENERAL LINGUISTICS

that one should proceed from ·poti- 'chief' and that the value of Lith. pats '(him)self' resulted from an appositional use, which Meillet did not go on to explain;8 this hypothesis is hardly compatible with the obvious antiquity of the particle. More likely, but not exempt from difficulties, is the opinion of H. Pedersen, who takes the meaning of "master" from "self," citing not exact proofs but parallels: he compares certain uses of "himself" to designate the "master of the house," such as Gr. autOs, Lat. ipse, Dan. dial. han selv 'the mistress of the house,' Russ. sam, sama 'barin i barynja.'9 But all that these examples can prove is that in a very special situation, as is that of familiars or domestics, a pronoun suffices to refer to the personage of authority. Slaves in Greek or Latin comedy express themselves thus, but not free men in the solemn language of religion or poetry. The use of ipse for the master of the house is a simple fact of "parole," and it has never reached the level of "langue." Besides, it is too sporadic and recent to account for the obviously archaic and "noble" forms such as the Sanskrit pair patilpatni, and the Greek, posis/potnia. It has not been observed that this "ancillary" use of autos, ipse, etc., has ever produced a lexical denomination for "master" as such, or a derivation starting with this sense. In short, these parallels are both too limited in their sphere and too "familiar" in style for them to be seen as other than "situational variants." The pronouns ipse and autos can occasionally designate the master; they have never signified "master" outside their context. They do not assist us in recovering the connection between the two forms of ·pot(i)-. The manner in which the forms of each series are distributed respectively is worthy of observation. It will be noted that Hittite, an archaic dialect in many respects, possesses only the particle -pet 'self' (apas-pet 'himself, he exactly'), and there is no trace of a nominal form like ·pot(i)-. This makes one assume that ·pot(i)- has a chance of being secondary. On the other hand, the nominal forms of the "master" group are not linked to any verbal root; when there is a verbal form, such as Sans.patyate, Lat. potior, it is clearly denominative. This is, therefore, a lexical family which is entirely and exclusively nominal. The terms under consideration are thus on the one hand a particle, on the other, a nominal form. I t is necessary first to state precisely the function of the particle -pet. There are two distinct expressions of equivalence in the Indo-European languages, which can be illustrated by the example of Gothic, which possesses both sama. and silba; by sama 'same' the equivalence is announced as a permanent fact of the object recognized in different aspects or in different instances; by silba 'self' the equivalence is opposed to otherness: "himself" to the exclusion of all others. It can be mentioned in passing that the value of emphasis and of contrast inherent in the expression of the "self" category

Semantic Problems in Reconstruction

261

leads to its being signalled either by reference to the corporal being (hence, Ind.-Ir. tanii-, Hitt. tuekka-, O.H.G. leip, Fr. en personne, en chair et en os, etc.), or by an emphatic denomination like the superlative (hence Germ. selbst, Gr. aut6tatos, Lat. ipsissimus [ef. met-ipsimus> O.F. medisme, Fr. meme] , Slav. sam as the superlative, etc.) as "exemplary" personifications of the notion. The function of the enclitic -pat in Hittite, -pat in Lithuanian, whose use is inherited, obviously answers to the notion of "self": Hitt. apas-pat 'that one exactly, himself,' Lith. ten-pat 'the self-same place,' as pats 'myself,' with a superlative value developed in Lithuanian: pats pirmasis 'the foremost.' In this function, the particle is attached to the pronoun, and then there occurs a selective association, which appears clearly in Iranian, in which -pati is incorporated into the reflexive-Avo xvae-paiti- 'oneself,' and especially the derived xvaepai()ya-, O.Pers. (h)uvaipaSiya 'one's very own,' in the O.Pers. predicative construction (h)uvaipasiyam kar- 'proprium facere, to appropriate' -for any person whatsoever, but always for one person. From this use can be deduced the explanation of the nominal *pet/pot, suffixed and nominalized by -i in *poti, which means the person in his own right, the "ipse" with some sort of determination. Actually the derivative of pati-, Sans. patya-, constructed with the dative, preserves the sense of "to be proper to": asutis cdrur mddaya patyate, lit. 'an agreeable beverage is proper to drunkenness' (R. V. 8. I. 26), and Av. pai8ya- signifies 'to have in one's own right' (and not 'to be master of). This definition of *poti as 'the ipse, the being in person' is conditioned by the determination that actually always accompanies the term in the most ancient expressions: the *dems poti (Av. dJng pati-, Ved. dam-pati, Gr. des-p6tes) is literally 'the ipse of the house, the being itself of the family,' the one who personifies the social cell. That is what we transpose into the terms of our own culture by the usual translation 'master of the house.' From this proceed all the other compounds, which gradually come to include more territory: Sans. vis-pati, Av. vis-paiti-, Lith. vieS-pats 'the one who is the ipse of the *wik- = master of the clan,' etc. We find two indications that corroborate this interpretation. The sense of Lat. hospes (*ghos-pet-), which designates the one who receives as well as the one who offers hospitality, is better explained as the "ipse" rather than the "master" of the reciprocal benefit designated by *ghos(ti)-, in which the two members are equal partners. In addition, it now becomes possible to connect the series of compounds in *-poti to a formation of the same meaning, but of different structure, which belongs to western Indo-European. Saussure once called attention to the curious formation of the following terms: Lat. dominus tribunus, Goth. piudans 'king,' kindins 'hegemon,' O.Ice. drottenn 'prince', they are all secondary derivatives in *-no- of base terms for designating chiefs:

PROBLEMS IN GENERAL LINGUISTICS

the dominus (*domo-no-) is the chief of the domus as the piudans (*teuta-no) is of the piuda. 10 If we compare the series of derivatives in -no- with that of the compounds in -poti, we shall see that they are parallel and share in common elements: *domo-no and *dem(s)poti-; *genti-no- (Goth. kindins) and *gentu-poti (Av. zantu-pati); a Latin *vicinus would be the counterpart of Sans. vii-patio This correlation between the terms in -no- of the western domain and the compounds in -poti, which abound especially in Indo-Iranian, suggests that they express the same notion. Now a derivative in -no- can hardly convey by itself the specific sense of "chief, master"; *domo-no-, *genti-no-, must simply signify 'the one of the domus, the one of the gens,' that is, in fact, the one who personifies it and in some way assumes it, who acts in its name and has authority over it. That is precisely the value which *poti indicates by itself: the representative personage, an ipse invested with authority in the social unit, whom we call the "master." If this is so, the base of the semantic history of *poti 'master' is to be found in the syntagms or in the compounds of which *poti is the second member. This is what the facts confirm: Sans. pati- 'master' in the free state is extracted from compounds in which it acquired its meaning. But then, what about the special acceptation of the term, that of 'husband,' attested to by Sans. patiand Gr. posis? Is it only the husband as "master" of the wife? This would satisfy a simplistic conception of Indo-European conjugality, but it would be belied by the feminine patni, potnia. The denomination is undoubtedly related to ancient usages concerning which one of the compounds, Goth. brup-faps, informs us indirectly. The relation of brup-faps 'numphios, bridegroom' to brufis 'numphe' is illuminated by the modern forms Brautigam, bridegroom (for *-goom), O.E. bryd-guma, in which -faps was replaced by the term for 'man' (-guma) in order to indicate "the man of the bride," that is, "the masculine partner of the bruti. " We must refer here to the very ancient formulae in which the future spouses were placed opposite one another as partners in an alliance; in Rome, ubi tu Gaius, ego Gaia; in India, amo 'ham asmi sd tvam 'I am he here; you are she there.'ll The same thing is true in this case; the pati and the patni, the posis and the potnia (-poina) are properly the 'ipse' and the 'ipsa' of the engagement that unites them. That is why the masculine partner of the bruti is denominated as *bhriiti-poti-, in which *-poti has the same function as -pet- in Lat. hospes. In this reconstruction the nominalization of the particle pet/pot- in -poti and the use of the particle with a pronoun in order to emphasize ipse-ness can be seen as the decisive factors in the semantic history of the two morphemes to be identified. The development of the syntagms (*dems poti) and the compounds is linked to the institutional value of the designations thus created in the very structure of Indo-European society. The man qualified

Semantic Problems in Reconstruction

by a title in ifE-poti is originally not the chief or the master but the representative of the social division. The Latin data deserve to be considered in their totality because they represent, in their variety of meanings and syntactic functions, a summing up of the entire process. The importance assumed in Latin by the group of posse, potens, potentia, potestas and the predominance of the notion of "power" in the modern derivatives have blinded philologists and linguists to the correlations that articulate this whole semantic group and, in particular, the conditions under which this notion of "power" was formed. As a point of departure, we find the inheritance of the enclitic particle (mea)pte, which serves to stress what is peculiar to one, ipse-ness: suopte pro suo ipsius, ut meopte meo ipsius, tuopte tuo ipsius (P. Festus 409. I). It will be noted that utpote signifies not 'as is possible,' but 'as is proper (to the circumstance), as is natural,' and that the sense of the comparative adverb potius 'rather, preferably,' super!. potissimum 'especially,' leads to a pote 'exactly, precisely, properly,' like the Hitt. -pat above. 12 The value of "that which is proper" is thereby introduced into the nominal forms, stressing possession as "property" [Fr. propriete, adj. propre, e.g., what is properly and peculiarly one's own]. In effect, compos signifies literally 'who is put in possession of,' not only in compos sui (or mentis, animi) 'who is in possession of his wits' or, as in French, "maitre de soi," but also in compos culpae (PI. True. 835) 'who is in possession of his fault, who identifies with it, who assumes the responsibility for it,' compos voti 'who has the possession of his wish, who has made it his own' (= who sees it realized), in manifest connection with the value of the Avestan compound xvaipazOya- 'proprius.' Such is obviously the sense of poti- inpossideo, lit. 'to occupy as one's own.' From the 'ipse' to the derivative 'proprius' a relation can be traced that fixes the sense of "possession." The archaic present potio signifies 'proprium facere, to make something the proper possession of someone': eum nunc potivit pater servitutis, lit. 'his father made him the possession of slavery' [i.e., made him a slave] (PI. Amph. 177). One must also take into account the decisive fact that potis tends to be constructed predicatively; thus we see how potis sum facere, lit. 'I am the selfsame one to do-je suis a meme de faire-ipse sum qui faciam,' becomes 'I can do.' Such is the makeup of the notion of "power," a "power" that depends on the distinctive quality of the person, on his self-identity, and not on human nature or the course of circumstances. It is the last stage of the process which leads from a particle of identity to the creation of a distinct nominal group, important and productive, and which Indo-European as well as Latin uses permit us to reconstruct with a fair degree of probability. In these analyses, which aim above all to illustrate some simple rules of methodology, we have chosen various examples. The problems considered

PROBLEMS IN GENERAL LINGUISTICS

are of varying complexity and of different levels, situated in the synchrony of the same language or in a perspective going back by degrees to a distant prehistory. They were chosen for their value as types and because it seemed to us that each of them could be brought to its solution. A methodology at grips with the difficulties of a real problem at least lets itself be judged by the solutions it proposes, whereas by reasoning from established conclusions one is sure of winning without risk and of teaching only what is known. In all the cases discussed a problem of relations was involved, and it is by relations that a semantic structure is defined. The alert reader will undoubtedly discern in the procedure followed here the same preoccupations that come to light in other parts of present-day linguistics, and even certain analogies in the object of the research. The preceding considerations turn upon a single question, which is the identification of distinctive features as opposed to variants: how to define the distribution and combinatory capacities of a "meaning"; how a sense taken as different from another can simply represent one of its variants; how the variant of a meaning is "semanticized" in its turn and becomes a distinct unit-these are all problems which could be immediately transposed into terms of phonemics. But semantic notions, which are more complex, more difficult to objectify and especially to formalize, since they are entangled in the extralinguistic "substance," call first for a description of those uses which alone will allow a meaning to be defined. And this description itself requires that we free ourselves of false evidence, of references to "universal" semantic categories, and of confusions between the data to be studied and those of the language of the describer. It is perhaps in the work of reconstruction that these conditions are the most severe.

From Word 10 (1954) : 251-264

TWENTY -FIVE

Euphemisms Ancient and Modern

THERE IS SOMETHING PECULIAR and paradoxical in the universally accepted explanation of the Greek term that expresses "euphemism."l The dictionaries give to E'lXpr;/lEtV two opposing meanings, and the one that is given first contradicts what it means: "to speak words which augur weW' and, consequently, "to avoid words which augur ill," hence, "to maintain silence." Thus, according to the literal definition of Liddell-Scott-Jones: "avoid all unlucky words during sacred rites: hence, as the surest mode of avoiding them, keep a religious silence." But the second meaning is the opposite: uto shout in triumph." This amounts to establishing a euphemism for a euphemism. But neither the actual meaning nor the historical uses of the word agree with this pseudological schema. In order to see the impossibility, it is enough to observe that the two meanings are found in the same authors; so that if we have to admit "to be silent" as the first meaning, that of uto cry out" becomes incomprehensible; and furthermore, E'I)(pr;/l{u, EVcpr;/lta/loc;, which were already in common use among the Greeks in the sense of "euphemism," cannot be related to either the one or the other. A confusion between the values of langue and those of parole' (in the Saussurian sense) has been introduced into the interpretation of these words. The religious acceptations, with all their overtones, associations, and overlappings are related to uparole." But these acceptations can only be determined by beginning with a purely linguistic value. In the study of the vocabulary of religion, as with all special vocabularies, it is necessary to separate the two aspects of the problem if we wish to understand the nature of the actions that are involved in it. We have then to begin by restoring the proper significance of Evcpr;pElv, EVcpr;/l{U, and there is no question but that this is positive. Because this obvious point has not been recognized, it is necessary to assert that EVcpr;/lE IV always means 'to utter words which are auspicious,' and only this. To mention only a few examples as proof, the compound brEVcpr;pElV is found as early as in Homer, where it clearly means 'to assent by an auspicious outcry' (naviec; enEVcpr;pr;aav ~ AX atol [Il. 1. 22.

PROBLEMS IN GENERAL LINGUISTICS

376]) and often in poetry c1)(Pwu ;iv occurs in the sense of 'to utter auspicious outcries' (Aesch. Ag. 596; Eum. 1035); xiAa~o~ r;Vcp~/l/rJC1ev (Aesch., Pers., 389); oAoAvy,ud~ evcpr;,uwv (Ag. 28); or in actual discourse, evcpr;,uov br:o~ (Suppl. 512); evcpr;,uo~ ,uovoa (Suppl. 694); avotyuv evcpr;,uov oTo,ua (Ar. Birds 1719); AOYWV evcpr;,uta (Eur. lA. 1469), etc. How then did the negative sense originate? A passage in Herodotus (3. 38) is helpful in understanding this. Darius asked some Indians what price they would accept for burning their deceased fathers: Ot ~i a,uf3woaV1:e~ ,uiya evcpr;,uietv ,utV bdAevov 'ceux-ci se recrierent fort et prierent Darius de ne pas prononcer des paroles de mauvais augure [they protested vehemently and begged Darius not to speak such ill-omened words), (Legrand). The locution evcpr;pielV ,ulV bdAeVOV shows that the verb preserved its basic meaning but that the circumstance in which it was spoken as an appeal in the form of a prohibition confers on it for us the negative meaning, 'do not speak of misfortune!' The concern here is to reverse the effect of evoking something unlucky. This acceptation results entirely from a context in which the verb is introduced in the form of an appeal to the evcpr;flta in order to combat the effect of words that were considered improper and which might bring down misfortune. Indeed, this use of Evcpr;,ueiv is always found in the imperative or in substitutes for the imperative, and also as an invitation to protect by words (cf. Lat. favete linguis) the course of a ceremony that even innocuous words would disturb: Evcpr;flfjoat XiAEafJe (II. 9. 171, the only Homeric example); eVcp1},uet, Evcpr;,ueiTE (Ar. Clouds 297; Ach. 241); eVcp'Y},uOV xol,ur;oov oTo,ua (Aesch. Ag. 1247); YAwooav evcpr;,uov cpieElV (Choeph. 581 ); evcpr;,uo~ 'LoOt (Soph. Fr. 478), etc. That, in practice, this injunction became the equivalent of "be silent!" changes nothing in the meaning of the verb. There is no instance of evcpr;,ueiv, meaning 'to keep silent,' being freely used in the sense of (]twnuv in a narrative context, but only occasions in the religious service when the invitation to "speak auspiciously," issued by the herald, obliged the worshippers to cease all other talk. The effect of the religious use upon the meaning of the word is obvious. It is necessary in order to appreciate a euphemism to recreate as far as possible the conditions of its use in spoken discourse. An expression like e'L Tl n&'Oolflt, ijv Tl n&.()w 'if something happens to me (i.e., if I die)' obviously does not authorize positing na()eiv Tl in the sense of "die." The situation alone determines the euphemism. And that situation, depending on whether it is permanent or occasional, modifies the type of the euphemistic expression according to norms peculiar to each language. Everything hinges on the nature of the idea that one wishes to bring to mind while avoiding naming it specifically. If the idea is among those which the social and moral norms condemn, the euphemism will not last; it will be

Euphemisms Ancient and Modern

contaminated in its turn and will have to be regenerated. It takes some thought to discern terms which were originally "respectable" in the Lat. meretrix (cf. mereor 'to earn'), Gr. 7l'X)(!'V'Y} 'whore' (cf. 'llE(!'V'Y}IU 'to buy'), Goth. hors ''ll6(!'Vo~, !1otx6~' (cf. Lat. carus 'dear'). But other ideas are only occasionally unfavorable, and the expression, according to the individual case, will be direct or will receive a substitute. For instance, in Avestan, the opposition of "white" and "black" is normally expressed by the adjectives auruIa- and sama(syama-, syava-). It is used in the symbolic figures in the mythology of the creation of opposing forces; the star Tistriya takes the form of a white horse (auruIa-),his enemy, the demonApaosa, a black horse (sama-), cf. Yt. 8. 20-21. But the same text (8. 58) prescribes offering to Tistriya 'a white sheep, or a black one, or one of any color as long as it is uniform' pasiim auruInn va vohu-gaonam va. This time the offering is consecrated to Tistriya, and nothing which is offered him should evoke the world of the daevas, so "black" is called vohu-gaona- 'of a good color' in order to exorcise sama-. 2 Sometimes an expression, which has become so commonplace that it no longer rouses the attention is illuminated by the beliefs associated with the idea it expresses. Those who say, as in French, "de bonne heure" [in good time, i.e. early] for "tot" [early] (cf. zu guter Zeit), are no longer aware of the peculiarity, which is nevertheless real, of Lat. mane 'early' being the adverb from manus 'good, favorable.' There is not yet a satisfactory explanation for this link between the idea of "early" and that of "good." For to bring up matiitinus, matiirus, as does J. B. Hofmann (in A. Walde, Lateinisches Etymologisches Wiirterbuch, rev. ed. [Heidelberg, 1965], 2: 25) in order to justify an original sense of "rechtzeitig," is to discount the religious value of manus while leaving the essential point obscure: just why was the morning so characterized? We must take into account ancient concepts that are still reflected in the Roman calendar. The days were not simply apportioned as fasti and nefasti. There were in addition divisions within certain days. It is known through Varro that there were certain dies ftssi that were unpropitious in the morning and propitious the rest of the time, and there were dies intercisi, unpropitious in the morning and evening and propitious during the interval. Morning thus had a special quality that predisposed it to prohibition. On this same subject, there is also extremely interesting evidence that comes from another people. E. Destaing collected from among the Berbers, at the dictation of an educated native speaker, what amounts to a treatise of linguistic taboos. 3 Among the very specific indications as to what brings about the use of euphemisms, there is one that recurs in connection with almost all the names of animals, tools, etc.; it is in the mornings that they are under the strictest prohibitions. "Experience has shown that the unpropitious influence of creatures and things, as well as that of the words which designate them, is

PROBLEMS IN GENERAL LINGUISTICS

exerted primarily in the morning. Consequently there is a whole category of taboo words which are forbidden only in the morning, before the meal taken in the middle of the day. This is the case with the names for broom, needle, cooking-pot, etc."4 Among other euphemisms that in Berber are kept for the language of the morning, the one which concerns hares is worth noting. Instead of calling the hare autUl, they say bu tmezgin 'the animal with long ears.' This immediately recalls the Indo-European expressions, Gr. Aayw6~ 'animal with hanging ears,' Pers. xargos 'animal with donkey's ears,' which must also be substitutes. 5 The Berbers are so aware of the portents of morning that if a man leaving his house at the beginning of the day sees a needle on the ground, "he picks it up, throws it far away, and angrily returns home in order to change his morning. How can he change his morning? He enters the house, goes to bed, shuts his eyes, and pretends to sleep for a minute. Then he goes about his business. Or he will take the utensils with which dinner was served the night before, and if there is any food remaining, he will eat several mouthfuls. If there is nothing cooked, he will take some flour, put it in his mouth, and go off saying, 'This is the real morning and not the other!' "6 Morning is, in effect, the dangerous time when, at the departure of night, the fortunes of the day are decided, whether good or evil. The Latin expression mane must be derived from this belief, from which one can now recognize the same euphemism in the adjective man is applied to the spirits of the dead, to the manes. Just as fearsome spirits are propitiated by their being called "good," so the beginning of the morning can be made favorable by characterizing it as the "good hour," or mane. This is a novel example of the process known through Gr. Ev#cv{~c~. In all these examples it is a matter of a fixed notion whose religious value is a constant. It receives a stable designation which also always belongs to the sacred vocabulary. The process consists in endowing an unpropitious notion with a propitious name. But for other ideas there is also a different process by which the expression considered bad is desacralized by substituting for it an equivalent which is remote or much weakened. This latter process can account for the various and not always fully understood ways for saying "kill" in Greek. One of these merits particular mention. Herodotus uses 'Karax(]ua()m for 'kill,' along with ano'Krctvat, but although each of the two verbs seems to be freely used for the other in the course of the same account, the usage is conditioned by reasons resulting from the circumstances. Astyages, in order to get rid of his daughter's son, who, according to a prophecy, was to dispossess him of his throne, orders Harpages to take him away and kill him: CPE(]WV ~e l~ acwvrov dn6uruvov (I. 108). The order is transmitted by Harpages to

Euphemisms Ancient and Modern

26 9

Mithridates in the same brutal form: xat/uv' AaTvay1j~ EvTe}.).. sTat anoxTE ivat

(I. II I). But in order to induce Mithridates to committhe murder, Harpages threatens him personally with the worst of deaths in the event that he should disobey: oJ-t(}eq> rep xaxtaup as l5taxe~asa(}at (I. 110). When Astyages discovers later that his order has not been carried out, he makes Harpages corne to him and asks: 'By what means did you bring about the death of the child born to my daughter whom I had turned over to you?' Teq> l5~ floecp TOV nail5a xaTsxe~aao TOV TOt naetl5wxa sx (}vyaTeo~ ysyoVOTa TPj~ sfl'f;~; (I. 117). It can be seen that l5taxeua(}at is used as a mitigation of anOXTstVat and that it appears in speech as a less explicit expression. In another passage (3- 36), Cambyses orders his guards to seize Croesus and kill him: ).af3ovTa~ fltV unoxTs ivat. But they prudently hide Croesus so that if Cambyses should change his mind, they would be rewarded; and if not, there would always be time enough to put him to death, TOTS xataxe~aaa(}at. The historian is interpreting the thoughts of those to whom this killing was repugnant. Here is another example of this same contrast between the brutality of an idea formulated in a decision and a less explicit expression at the moment of carrying it out: the Lacedemonians had decided to kill Minyans, ToiO't dw Liuxsl5atflov{otO't ll5o~s aVTov~ anoxTsivat; but at the moment of executing them ... , snst wv lfls),),ov acpEa~ xaTaxe~asa(}at (3. 146). This verb occurs again in a request for punishment and in reproducing the terms of the request: lnsflnov enEte1jaofleVOv~ cl xaTaxe~aWVTat T~V vnoC axoeov T WV (}EWV 'they sent to ask the oracle if they should execute the subpriestess of the goddesses [who had delivered secrets to Miltiades], (6. 135); Ot' E).awvaLOI, Tip IIeWTsa{).sq> ilflOeeo'Vis~ el5eovTo fltV XaTaXe1ja(}ijvat 'the Eleontines, in order to avenge Prothesilaus, had asked that he be put to death' (9. 120). Finally, Herodotus uses xaTaxeija(}at with the reflexive for 'to kill oneself': ).tyovat . .. aVTov fltV b TriO't (}vetnat xaTaxe~aaa(}at ewvTov (I. 82); in the same sense one also finds aViov l5taxeua(}at (I. 24) and ewvTov xaTseyaCea(}at (ibid.). It appears then that xaTaxeua(}at, &axe ua() at, xaTeeyaCea()at, are euphemistic expressions meaning 'to finish someone off, to liquidate him' in cases where feeling prohibits a cruder expression. The range of the usages explains and justifies the semantic deviation. The French executer [to carry out] in the sense of 'to put to death' is in conformity with this same feeling. This acceptation proceeds from the official euphemism executer (a mort) [to carry out a sentence (to death)] and from the one which designates the hangman as executeur de fa haute Justice, des hautes oeuvres [the executor of (the one who carries out) supreme justice, supreme deeds] (cf. Germ. Schar/ric/lter). The discredit attached to the function of the hangman caused him to be designated by various euphemisms in Greek: l5~flLO~ (scil. l5ov).o~), XOtvo~ l5~flw~ (Plato Laws 87 2b ), l51jflOXflLVO,

PROBLEMS IN GENERAL LINGUISTICS

(Soph. fro 780; Antipho 1. 20; Iso. 17. IS)' In Latin, on the other hand, a name which was an insult was preferred: carnufex. But what exactly does carnufex mean? The literal sense is just that which Donatus (Hec. 441) defines: earnifices dieti quod carnes ex homine faciant. This compound nevertheless has something peculiar about it when it is compared to opi-fex, auri-fex, arti-fex, etc. It gives the impression of being a translation. And it is indeed as a translation that it can be explained: carnu-fex is an exact calque of Gr. u(!EOv(!y6~ 'butcher,' already found in Aesch., U(!EOV(!YOV iiftu(! (Ag. 1592); cf. KuruKf!EOeYE tV 'to put into pieces' (Hdt. 7. 18 I ); U(!EOV(!yr;lJov lJwanuaavr E~ rov~ avlJ(!a~ 'dismembering them limb from limb like butchers' (Hdt. 3. 13). Latin thus changed the Greek word for "butcher" into a term designating the "hangman" (and this, in spite of everything, is a kind of euphemism), while at the same time the word maeellarius, derived from macellum, which is also from the Greek, is kept for the word "butcher." In a completely different area, Havers has rightly emphasized the euphemistic character of expressions for "to extinguish the fire" in relation to popular beliefs concerning fire as a living being. 7 To all the evidence that he has brought together might be added a few Iranian examples. A very powerful superstition in Iran and Afghanistan forbids extinguishing a flame by blowing on it. 8 This does not indicate that one cannot properly say "to extinguish the fire"; in fact, there is an even stronger expression, alas kustan 'to kill the fire' (cf. Sans. pari-han- in the same sense). But in common usage a euphemism prevails: sakit kardan 'to appease' and, especially, xamiis kardan 'to make silent, to quiet (the fire),' or Tuxsat dadan 'to give leave to it'; of the fire itself, ruxsat sude 'it has taken leave, it is extinguished' is said. In Afghanistan, the ordinary locution is (ataS) gul kardan (cf. Hindi: gul karna) 'to extinguish,' passive, gul sudan, which is also a euphemism but one in which the sense of gul is not completely clear.9 This is probably the word that the older dictionaries interpret as "the snuff of a lamp or a candle," and the expression would mean something like "to snuff the flame." The aim of all these processes is not simply to mitigate the idea of "extinguish." Just as in the Vedic rituai of sacrifice, the victim is 'appeased' (Samayati) or 'made to consent' (sa1Jljiiapayati), when in actual fact it is "strangled," so the fire that is extinguished is "appeased." All this recalls Lat. ignem tutare, which is indeed to be understood as 'to calm, to appease (the fire),'lO and this confirms the euphemistic origin of Fr. tuer [to kill].

From Die Sprache 1 (1949) : 116-122.

TWENTY-SIX

Gift and Exchange in the Indo-European Vocabulary

IT WAS THE VERY GREAT contribution of Marcel Mauss, in his now classic "Essai sur Ie don,"1 to have revealed the functional relationship between gift and exchange and to have defined thereby a whole group of religious, economic, and judicial phenomena belonging to archaic societies. He showed that the gift is only one element in a system of reciprocal prestations which are at once free and constraining, the freedom of the gift obliging the recipient to a countergift, which engenders a continuous flow of gifts offered and gifts given in return. Here is the principle of an exchange which, generalized not only between individuals but also between groups and classes, stimulates the circulation of wealth throughout the entire society. The game is determined by rules that become fixed in institutions of all sorts. A vast network of rites, celebrations, contracts, and rivalries organizes the mechanics of these transactions. The demonstration made by Mauss was based primarily upon archaic societies, which furnished him with a mass of conclusive evidence. If one seeks to verify this mechanism in ancient societies, particularly in the IndoEuropean world, convincing examples become much more rare. It is true that Mauss himself described "an archaic form of contract among the Thracians," and he also discovered in ancient India and Germany traces of analogous institutions. In addition, one must allow for chance discoveries, always possible in this vast domain in which the investigation has not been systematically pursued. The fact remains that these societies are much more difficult to explore and that, as far as usable documents are concerned, one cannot count on a large amount of sure and specific evidence, if one wishes it to be explicit. We do have nevertheless some less apparent facts, which are all the more valuable for not having run the risk of being distorted by conscious interpretations. These are the facts presented by the vocabulary of the IndoEuropean languages. One cannot use them without an elaboration based on the comparison of attested forms. But that comparison will result in conclusions

PROBLEMS IN GENERAL LINGUISTICS

which will supply to a rather large degree the absence of evidence for the most ancient periods of our societies. Several examples will be brought forth and analyzed in order to obtain whatever information they can offer about the prehistoric notions of gift and exchange. In most Indo-European languages, "to give" is expressed by a verb from the root *do- which also has a large number of nominal derivatives. There seemed to be no possible doubt about the constancy of this signification until it was established that the Hittite verb dii- meant not 'give' but 'take.' This caused considerable confusion, which still lasts. Should Hittite dii- be considered a different verb? We cannot assume this without misgivings. Must we, on the other hand, admit that the original sense of *do- was 'take,' faithfully preserved in Hittite dii- as well as in Indo-Iranian ii-dii- 'receive'? This would reverse the problem without making it any easier; it would remain to be explained how "give" could have come from "take." In reality the problem seems insoluble if we seek to derive "take" from "give" or "give" from "take." But the problem is wrongly put. We shall consider that *doproperly means neither 'take' nor 'give' but either the one or the other, depending on the construction. It must have been employed like English "take," which permits two opposed meanings: "to take something from someone" but also "to take something to someone, to deliver something to someone." Cf. also, "to betake oneself, to go"; besides, in Middle English, taken meant 'to deliver' as well as 'to take.' Similarly, *do- indicated only the fact of taking hold of something; only the syntax of the utterance differentiated it as 'to take hold of in order to keep' (= take) and 'to take hold of in order to offer' (= give). Each language made one of these acceptations prevail at the expense of the other in order to construct the antithetical and distinct expressions for "taking" and "giving." Accordingly, in Hittite dii- means 'take' and is opposed to pai- 'give,' while in most of the other languages it is *do- which means 'give,' and a different verb which assumes the meaning of "take." Some traces of the double possibility survive; even though the distribution was fixed in Indo-Iranian, the verb dii- 'to give,' with the preverb iiindicating movement toward the subject, means 'to receive.' It seems, then, that the most characteristic verb for "to give" was marked by a curious semantic ambivalence, the same sort of ambivalence affecting more technical expressions like 'buy' and 'sell' in Germanic (Germ. kaufen: verkaufen) or 'borrow' and 'lend' in Greek (~avel'w : ~avcl'o!wt). "To give" and "to take" thus proclaim themselves here, in a very ancient phase of Indo-European, as notions that were organically linked by their polarity and which were susceptible of the same expression. Now *do- is not the only example of this. For a long time there has been

Gift and Exchange in the Indo-European Vocabulary

273 a question about the etymology of the verb for "take" in Germanic: Goth. niman, Germ. nehmen, which assumes a root *nem-. One would naturally think of relating it to Gr. YEflW. Comparatists have always refused to do this, claiming that there was a difference in meaning. 2 But the meaning must be defined with some precision before it can be decided if it is really an obstacle to the relationship. The Greek verb yeflW has the two values of 'to give legally as an allotment' (Zd)~ yeflet oA{3OY ay(}ewnouJt rOd. 14. 188] and 'to have legally as an allotment' (noAty ~JEfleLV) [Hdt. I. 59]).3 In Gothic, niman does indeed mean 'to take' in various acceptations. But a compound of this verb is of special interest: it is arbi-numja 'heir,' lit. 'the one who takes (= receives) the inheritance.' Now the Greek term that arbi-numja translates is y.AYjeovoflO~ 'heir.' Is it chance that (x;'YjeO)YOflO~ and (arbi)numja are formed from yeflW in Greek and from niman in Gothic? Here we have hold of the missing link which allows us to join the meanings which history has separated. Goth. niman means 'to take,' not in the sense of 'to take hold of' (which is greipan, Germ. greifen) but in the sense of 'to receive' and more exactly, of 'to receive as an allotment, into possession,' which is precisely the same as one of the two acceptations of Gr. yeflw. The connection between YEp,W and niman is now restored, and is confirmed by the ambivalence of *nem-, which indicates legal attribution as given or as received. 4 Let us now turn to the very notion of "gift" in the form which is the most constant throughout most of the Indo-European languages. We observe that, in general, nominal forms derived from *d6- were used. Now, it happensand this fact has been barely noticed-that within a single language, several of these derivatives will be employed simultaneously, being differentiated by their suffixes. The coexistence of these "synonyms" should arouse attention and call for a strict verification, first because they are not synonyms and, more especially, because the simplicity of a notion such as "gift" would not seem to require multiple expressions. Ancient Greek had no fewer than five distinct and parallel words for "gift," *d6-, and our dictionaries and translations render them identically as 'gift, present' : {)wr;, {)o(]t~, {)weOY, {)weea, {)w ilvYj. 5 We must try to define each one of them specifically by virtue of its formation. The first, {)w~, has only one example, in Hesiod: c5d)~, aya(}ry, aena~, c5i y.ax1] 'to give is good, to ravish is evil' (Works 354); a root word which, like aena~, must have been an invention of the poet for an expression as simple and as little differentiated as possible for "gift." In c50ate the notion is presented as an effective accomplishment; it is the act of giving susceptible of being realized in a gift: 6 xat 01 c5oat~ Eaa87:at fa();'~ '(the one who will devote himself), we shall give him a precious gift' (Il. 10. 213)' This time, the gift is promised in advance, designated in detail, and is to recompense a bold deed. The next two, {)weo'V and c5weea

PROBLEMS IN GENERAL LINGUISTICS 274 must be taken together: the first, (jwQov, is indeed the gift of generosity, of gratitude, or of homage, which is incorporated into the object offered; and (jwQEa properly designates, as an abstraction, ·the providing of presents' (cf. Hdt. 3. 97) or the ·totality of presents' (ibid. 3. 84), whence the adverbial use (jWQEa-V ·in the manner of a present, gratuitously.' Aristotle defines (jwQsa precisely as a (j60't~ avan6(joTo~ (Top. 125a. 18), a (j60'l~ that does not impose the obligation of a gift in return. Finally there remains the most significant term, (jwrtvYj, which is also a gift but of a completely different sort. The (jwrtvYj, in Homer, is the obligatory gift offered to a chief whom one wishes to honor (Il. 9. 155· 297) or the gift that is due one as a guest; Ulysses, received by Polyphemus, feels he has a right to count on the (jwrtvYj, which is a part of the duties of hospitaJity: ei n n6Qot~ ~EtV~toV ~e Kat aAAw~ I (jolYj~ (jwTlvwv if iE I ~elvwv ()tflt~ EO'Tlv (Od. I I. 267). Alcinous, welcoming Ulysses at his home, does not wish to let him leave without having brought together the whole (jwrtvYj that is meant for him: El~ 0 XE niiO'av I (jwilvYjv n;AtO'w (Od. I I. 35 I). The uses of the word in Herodotus confirm this technical sense. A man, wishing to befriend the husband of a woman whom he desires, offers him as a (jwTlvYj any of his possessions that the husband might desire, but on condition of reciprocity (Hdt. 6. 62). One cannot emphasize more clearly the functional value of the owrtvYj, of this gift that obliges a gift in return. This is the invariable sense of the word in Herodotus; whether the (jwTlvYj is intended to call forth a gift in return or whether it serves to compensate for a previous gift, it always includes the idea of reciprocity: it is the gift that a city is compelled to give the person who has done it a service (I. 61); the gift sent to a people in order to engage their friendship (I. 69).7 Whence (jwnva~w (2. 180) ·to collect the (jwTivat' in the form of voluntary contributions from the cities towards a common work. In an inscription from Calauria, (jwrtvYj relates to the "rent" due in kind from one who has obtained a concession of land (I.G. 4. 841. II; third century B.C.). We have in (jwrtvYj the notion of a gift in return or a gift which calls for a return. The mechanism of the reciprocity of the gift is revealed by its very meaning and is related to a system of offerings of homage or hospitality. Up to this point we have considered words whose sense brought them to our attention immediately. But a valid inquiry must and can go well beyond the terms that have an explicit reference to the gift. There are some which are less apparent, not immediately obvious, and which sometimes can be recognized only by certain particular qualities in the meaning. Others preserve their proper value in only one part of the Indo-European domain. We must make use of both in order to reconstruct this complex prehistory. An obvious connection joins the notion of the gift to that of hospitality. But one must distinguish among the terms relating to hospitality. The ety-

Gift and Exchange in the Indo-European Vocabulary

275

mology of some of them, like Greek ~i'Vo; is not certain. The study of the word is thus involved with that of the institution and should be left to the historian of Hellenic society. More interesting are the terms whose evolution we can follow, ev~n, and perhaps especially, if this evolution has produced divergences in the meaning. One of these is the Latin word hostis. The term hostis will here be considered in its relation with other Latin words of the same family, which includes more than Latin (Goth. gasts, O.Slav. gosti 'guest/host'), but we shall put aside hospes, which cannot be analyzed with any certitude although it is certainly related. Well-known Latin evidence assists in the reconstruction of the history of hostis in Rome. The word still means 'foreigner' in the Law of the XII Tables, and this sense was familiar to Roman scholars. Varro (L.L. 5· 3) states: "hostis ... tum eo verbo dicebant peregrinum qui suis legibus uteretur, nunc dicunt eum quem dicebant perduellionem." And Festus (flf. 37) gives us this important definition in addition: II • • • ab antiquis hostes appelabantur quod erant pari iure cum populo Romano atque hostire ponebatur pro aequare." There are actually a series of proofs that hostire did indeed signify aequare, Several derivatiW!s confirm it, some of which are related to material operations. others to judicial or religious institutions. In Festus himself, we find redhostire 'referre gratiam,' and in Plautus: promitto . . . hostire contra ut merueris 'I promise to pay you back as you deserve' (Asin. 377). In addition, hostimentum is defined as 'beneficii pensatio' and 'aequamentum' (Nonius 3. 26) and, according to a gloss, more precisely, 'hostimentum dicitur lapis quo pondus exaequatur' (C.G.L. 5. 20g. 3). This meaning appears in Plautus, where it indicates the "compensation" for work and wages: 'par pari datum hostimentumst, opera pro pecunia' (Asin. 172). The same notion is present in hostus, which Varro specifies as a rural term: 'hostum vocant quod ex uno facto olei reficitur: factum dicunt quod uno tempore conficiunt' (R.R. I. 2f. 3); the sense is properly 'compensation, that oil which is obtained as a compensation for one pressing.' Hostorium was the name for the stick used to level the bushel (lignum quo modius aequatur, Prise. 2. 215. 17; C.G.L. 5. 503. 36). Augustine (Civ. Dei. f. 8) mentions a dea Hostilina who was in charge of equalizing the ears of grain (or perhaps of equalizing the harvest with the work realized). These clear and concordant indications are not diminished by certain glosses in the abridgement of Festus and Nonius, according to which hostire would mean 'ferire, comprimere, caedere'; this sense was deduced from archaic citations that were not exactly understood and which moreover refute it: in hostio ferociam (Pacuvius) and hostit voluntatem tuam (Naevius), the verb does not mean 'destroy' but 'compensate" counterbalance. ' An important term for this family is gained by annexing to it the word

PROBLEMS IN GENERAL LINGUISTICS

hostia. Hostia does not designate any offered victim at all but only the one which was intended to 'compensate for' the anger of the gods. Just as important in another domain is the term hostis, whose relation to all the others which surround it can be seen. The primary meaning of hostis is indeed the one Festus gives it: not just any "foreigner" but the foreigner who is pari iure cum populo Romano. H astis thereby takes on the meaning of both 'foreigner' and 'guest.' The equal rights that he enjoyed with respect to the Roman citizen were connected with his status as a guest. Hostis is properly one who compensates and enjoys compensation, one who obtains from Rome the counterpart of the advantages which he has in his own country and the equivalent of which he owes in his turn to the person whom he pays reciprocally. This old relationship was weakened, then abolished, as the status of civis came to be more rigorously defined and the civitas became the sole and ever stricter norm of judicial participation in the Roman community. The relationships regulated by personal or family agreements were wiped out in the face of rules and duties imposed by the state; hostis then became the 'foreigner' and then the 'public enemy' by a change in meaning that is connected with the political and judicial history of the Roman state. Through hostis and the related terms in early Latin we can discern a certain type of compensatory offering that is the basis of the notion of "hospitality" in the Latin, Germanic, and Slavic societies; equality of status transposes into law the parity between persons confirmed by reciprocal gifts. In order to approach a different aspect of the same notions, we shall resort to another Latin word whose meaning has been more stable but also more complex. An entire Indo-European phenomenology of "exchange," of which fragments survive in the numerous forms derived from the root *mei-, might be traced through and around munus. We should study in particular the Indo-Iranian notion of mitra, the contract and the god of the contract, a term whose authentic meaning largely overlaps that of the "contract." It is the equivalent in the human world of what the ria is in the cosmic world, that is, the principle of total reciprocity that bases human society on rights and obligations to the point that the same expression (Sans. druh-, Av. drug-) indicates the one who violates the mitra and the one who transgresses the rta. This profound and rich expression takes on a particular acceptation in Lat. munus. In literary use, munus means 'function, office,' or 'obligation' or 'task' or 'favor' or, finally, 'public spectacle, gladiatorial contest,' all acceptations relating to the social sphere. The formation of munus is characteristic in this regard; it contains the suffix *nes- which, as Meillet correctly observed, is attached to designations of a social or judicial nature (cf. pignus, fenus, funus, facinus). The unity of meanings in munus is found in the notion of respects paid or service accomplished, and this itself goes back to what Festus de-

Gift and Exchange in the Indo-European Vocabulary

277 fined as a donum quod officii causa datur. In accepting a munus, one contracts an obligation to repay it publicly by a distribution of favors or privileges, or by holding garnes, etc. The word contains the double value of a charge conferred as a distinction and of donations imposed in return. Here is the basis of "community," since communis signifies literally 'one who shares in the munia or munera'; each member of the group is compelled to give in the same proportion as he receives. Charges and privileges are the two faces of the same thing, and this alternation constitutes the community. An "exchange" which is constituted of "gifts" accepted and returned is something quite different from utilitarian commerce. It must be generous in order to be judged profitable. When one gives, he must give the most precious thing he has. This is what can be learned from certain terms that are etymologically of the same family as munus: O.Irish main, moin, which means 'present' and 'precious thing,' and especially Goth. maifims '~weov,' O.Ice. meidmar pI. 'jewels,' O.E. madum 'treasure, jewel.' It is worth noticing that Goth. maifims is not a gift in the sense that English "gift" would express. This word appears in the translation of Mark 7: I I, to render ~weov, but as the equivalent of the Hebrew word xoef3uv 'offering to the Treasure of the Temple.' The choice of maifims shows that in Gothic as in the other Germanic languages, the present of exchange must be of signal value. A comparison of vocabulary will reveal to us an institution analogous to the ones we have just discussed, but not so obvious. It is a type of donation almost abolished in historical societies and which we can only rediscover by interpreting the rather dissimilar significations of a group of words derived from *dap-: Lat. daps 'sacred banquet,' O.Ice. lafn 'sacrificial animal,' Arm. tawn 'feast,' Gr. ~anav17 'expense' (cf. ~anu.o 'break to pieces, consume, destroy'), and also Lat. damnum 'damage' (*dap-nom). The religious sense of some of these terms is clear. But in each of them the meaning has been narrowed down to only one particular aspect of a representation which goes beyond the sphere of the sacred and is realized in the domains of law and economy as well. As the nucleus of the meaning we shall set up the notion of "expense" as a manifestation both religious and social: a festive and sumptuous expense, an offering that consists of a large consumption of food, made for prestige and as a "pure loss." This definition seems to account for all the special acceptations arising from the fragmentation of an archaic conception. The Roman daps was a banquet offered to the gods, a real banquet with roast meat and wine which the participants ceremoniously consumed after having desacralized it. The antiquity of this rite can be seen in the formulae that consecrated it; according to Cato, these prayers were addressed to Jupiter: Jupiter dapalis,

PROBLEMS IN GENERAL LINGUISTICS

quod tibi fieri oportet, in domo familia mea culignam vini dapi, eius rei ergo macte hac illace dape pollucenda esto . . . Jupiter dapalis, macte illace dape pollucenda esto (Cato Agr. 132). The use of pollucere with daps emphasizes the magnificence of it: the verb always accompanies splendid consecrations in the ancient religious vocabulary. This can actually be seen in Ovid (Fasti 5. 515ff) when the poor peasant Hyrieus offers Jupiter, who is visiting him, a whole ox, his only possession, as a daps. Moreover, ancient derivatives of daps confirm the fact that this word implied largesse and associate it with festive banquets of hospitality: 'dapatice se acceptos dicebant antiqui, significantes magnifice, et dapaticum negotium amplum ac magnificum' (Festus). The verb dapinare, whether it is connected with daps or whether it is an adaptation of Gr. ~anava.v, signifies, in the only example of it that survives, 'to treat royally at the table': aeternum tibi dapinabo victum, si vera autumas (PI. Capt. 897). In Greek, ~an6.vrJ' of which, in general, only the commonplace acceptation of "expense" is retained, also implies largesse, an expense for display and prestige, although the term is no longer restricted to religious use. In Herodotus (2. 169), ~anavrJ signifies 'sumptuous ornamentation' in the decoration of a building. Pindar (Isthm. 4. 29) provides a significant use of it: llave).).6.vE:aat ~' B(!tCOpevot ~anavq. xai(!ov lnnwv '(the competitors in the games) in rivalry with the peoples of all Hellas took pleasure in expenditures on horses.' It really is, in effect, an expense of rivalry and prestige. If a new proof is necessary, it will be found in the sense of the adjective ~a1pt).~~ 'abundant, splendid,' which passed into Latin, in which dapsilis 'magnificent, sumptuous,' is associated secondarily with daps and renews an ancient etymological connection. The verb ~anava.v means 'to spend,' but it must be understood in a stronger sense: 'to spend' here means 'to consume, to destroy'; cf. ~anavrJ(!o~ 'prodigal, extravagant.' Hence, with the strict notion of a 'sacrifice with food' (Lat. daps, O.Ice. tafn) and of 'feast' (Arm. tawn) must be associated the idea of an ostentatious prodigality which is at the same time the consumption of food and the destruction of wealth. This clarifies the word damnum, so curiously separated from this semantic group. In damnum, there remains only the sense of 'damage suffered,' of material and especially pecuniary loss: it is the "expense" imposed upon someone and no longer consented to freely, the "loss" which is prejudicial and no longer a voluntary sacrifice; in short, a detriment or a penalty instead of a magnificent squandering. Jurists, who were also peasants, thus narrowed and reduced to a penalty what had been the sign of largesse and generosity. Whence damn are 'damno afficere, to impose a fine,' and in general, 'to condemn.' All these features help us perceive, in an Indo-European prehistory which is not so ancient, a socioreligious phenomenon of which we still retain many

Gift and Exchange in the Indo-European Vocabulary

279

traces in our vocabulary today. In English we say "to give a reception" and in French "offrir un banquet"; there are "expenses" of food and "sacrifices" of possessions made as social obligations and as fulfillments of the duty of hospitality. This analysis leads us finally to recognize, in the Indo-European word, the institution known as potlatch. It does not seem that the ancient classical societies knew that aggravated form of potlatch that several writers, Mauss in particular, have described among the K wakiutl or the Haida, or those extravagant challenges in which chiefs who were jealous of their prestige provoked one another to enormous destructions of wealth. But the fact still remains that the terms analyzed here refer to a custom of the same type as the potlatch. Although the theme of rivalry no longer appears, the essential features are really the same: the feast with an abundance of food, the expense which is purely ostentatious and intended to maintain rank, the festive banquet-all this would have no sense if those who had the profit of this largesse were not committed to requite it by the same means. Moreover, is it chance that the term potlatch is related in essence to offerings of food and means literally, Ito nourish, to consume'?8 Among all the varieties of potlatch, this must have been the most usual in societies in which the authority and the prestige of the chiefs were maintained by the largesse they distributed and from which they benefited in turn. It would be easy to extend these considerations further, either by pursuing the etymological relations of the terms examined, or, on the other hand, by studying the different Indo-European expressions for notions that are apparently identical. One example will show in what unpredictable form the notion of "exchange" may be revealed. As one might guess, "exchange" gives rise to a large vocabulary for specifying economic relations. But terms of this type have almost always been renewed, so that we must consider each language for itself. There is, however, one term which is at least fairly widespread in Indo-European and which is unvarying in meaning: it is the one that properly designates "value." It is represented by Gr. aA,g?avw, Sans. arh- Ito have worth, to be worthy' (cf. arhat Ideserving') also Av. arz-, Lth. alga Iprice, wages.' In Indo-Iranian and in Lithuanian, the sense appears to be rather general and abstract, not lending itself to a more precise determination. But in Greek, aAtpavw allows for a more exact interpretation than the dictionaries indicate in rendering it by 'to earn, to yield.' In Homer, aA,tpavw means indeed Ito get a profit,' but this sense is connected to a well-defined situation: the profit in question is the one that a captive brings to the man who sells him. It suffices to enumerate the Homeric examples. In order to move Achilles to pity, when he is ready to kill him, Lycaon implores him: Iyou once took me and led me to be sold at the market

280

PROBLEMS IN GENERAL LINGUISTICS

at Lemnos, where I brought you the price of a hundred oxen' ixaiofl{3otOV (JE iot ij;'qJov (Il. 21. 79). About a little slave who is offered for sale: 'he will bring you a thousand times his price' 0 (J'vfliv flV(!lov 6)VOV aAqJot (Od. 4· 453). Melantheus threatens to sell Eumaeus far from Ithaca 'so that he will bring me a good living' tva ItOt {3loTov nOAVv aAqJOt (Od. 17. 250), and the suitors invite Telemachus to sell his guests at the market in Sicily 'where they will bring you a good price' (J()eV xe iot a~tOv aAqJot (Od. 20. 383). There is no variation in the meaning of the verb and the full force of it is found in the epithet that describes maidens: nae()tvot a).qJeul{3oWt they 'bring in oxen' for their father who gives them in marriage. "Value" is characterized, in its ancient expression, as a "value of exchange" in the most material sense. It is the value of exchange that a human body possesses which is delivered up for a certain price. This "value" assumes its meaning for whoever disposes of a human body, whether it is a daughter to marry or a prisoner to sell. There we catch a glimpse, in at least one part of the Indo-European domain, of the very concrete origin of a notion connected to certain institutions in a society based on slavery.

From L'Annee sociologique, 3rd series,

2

(1948-1949):

7-20

TWENTY-SEVEN

The Notion of "Rhythm" in its Linguistic Expression

I T MIGHT BE THE TASK of a psychology of movements and gestures to make a parallel study of the terms that denote them and the psychological phenomena that they express, the meaning inherent in the terms, and the often very different mental associations that they awaken. The notion of "rhythm" is one of the ideas that affect a large portion of human activities. Perhaps it even serves to distinguish types of human behaviour, individual and collective, inasmuch as we are aware of durations and the repetitions that govern them, and also when, beyond the human sphere, we project a rhythm into things and events. This vast unification of man and nature under time, with its intervals and repetitions, has had as a condition the use of the word itself, the generalization, in the vocabulary of modern Western thought, of the term rhythm, which comes to us through Latin from Greek. In Greek itself, in which ev(}p,6~ does indeed designate rhythm, where does the notion come from and what does it properly mean? An identical answer is given by all the dictionaries: ev(}p,6~ is an abstract noun from ee tV 'to flow,' the sense of the word, according to Boisacq, having been borrowed from the regular movements of the waves of the sea. This is what was taught more than a century ago, at the beginnings of comparative grammar, and it is what is still being repeated. And what, really, could be more simple and satisfying? Man has learned the principles of things from nature, and the movement of the waves has given rise in his mind to the idea of rhythm, and that primordial discovery is inscribed in the term itself. There is no morphological difficulty in connecting evefl6~ to eEW by means of a derivation which we shall have to consider in detail. But the semantic connection that has been established between "rhythm" and "to flow" by the intermediary of the "regular movement of the waves" turns out to be impossible as soon as it is examined. It suffices to observe that EW and all its nominal derivatives (eevp,a, em], e6o~, evas, eVT6~) are exclusively indicative of the notion of 'to flow,' but that the sea does not "flow." ~Peiv is never said of the sea, and moreover, eV()fl6~ is never used for the movement of the waves.

e

PROBLEMS IN GENERAL LINGUISTICS

The terms which depict this movement are entirely different: aflnWTl~, eaxla, nAr;flV(!l~, aaAeVetv. Conversely, what flows, eeT, is the river or the stream, and a current of water does not have "rhythm." If eV()fl6~ means 'flux, flowing,' it is hard to see how it could have taken on the value proper to the word "rhythm." There is a contradiction of meaning between eeiv and ev () fl6~, and we cannot extricate ourselves from the difficulty by imagining -and this is a pure invention-that ev()fl6~ could have described the movement of the waves. What is more, eV()fl6~ in its most ancient uses never refers to flowing water, and it does not even mean "rhythm." This whole interpretation rests on the wrong premises. It is clearly necessary, in order to reconstruct a history which was less simple but which is also more instructive, to begin by establishing the authentic meaning of the word eV()fl6~ and by describing its use at its origins, which go very far back. It is absent from the Homeric poems. It is especially to be found in the Ionian authors and in lyric and tragic poetry, then in Attic prose, especially in the philosophers. 1 It is in the vocabulary of the ancient Ionian philosophy that we may apprehend the specific value of eV()fl6~, and most particularly among the creators of atomism, Leucippus and Democritus. These philosophers made eV()fl6~ (evaI16~)2 into a technical term, one of the key words of their teaching, and Aristotle, thanks to whom several citations from Democritus have come down to us, transmitted its exact meaning to us. According to him, the fundamental relationships among bodies are established by their mutual differences, and these differences come down to three: evafl6~, &a()ly~, y(]on~, which Aristotle interprets thus: ~tacpi(]eLV ya(] cpaat TO ov eval10 ",at ~ta()lYfi ",at T(]onfi' TOVTWP ~' 0 flev eval1o~ axiffla eaTlv, ~ ~e ~ta()ty~ Ta~t~, ~ ~e T(]On~ ()iat~. 'Things are differentiated by rlvafl6~, by ~ta()lY~, and by T(]On~; the evafl6~ is the axiffla ('form'), the ~ta()ly~ (,contact') is the Ta~t~

(,order'), and the T(]On~ ('turn') is the ()iat~ ('position') (Metaph. 98Sb 4). It is clear from this important passage that evafl6~ means 6xiffl a ('form'), which is confirmed by Aristotle by what follows in the passage, in an example borrowed from Leucippus. He illustrates these three notions by applying them respectively to the "form," "order," and "position" of the letters of the alphabet: 3 A differs from N by the axifl1a (or rlvafl6~), AN differs from NA by the Ta~t~, and I differs from H by the ()iat~. Let us hold on to the idea suggested by this passage that rlvafl6~ is the equivalent ofaxifl1a. Between A and N, the actual difference is one of "form" or "configuration": two of the strokes are identical-A-and only the third is different, being interior in A and exterior in N. And it is indeed in the sense of "form" that Democritus always uses rlV()fl6~.4 He wrote a treatise IIe(]l nov ~tacpe(]6vu))v rlvaflwv, which means 'on the variety of form (of

The Notion of Rhythm atoms).' His doctrine taught that water and air, eV()f1,iJ> ~tacpEeetV, differ from each other in the form that their constituent atoms take. Another citation from Democritus shows that he also applied ev()poc; to the form of institutions: oV~Epla P17Xavi] up vvv xa()EaTWTt ev()p0 pi] ovx a~LXE tV rov~ aexov-rac; 'there is no way, in the present form (of the constitution) to prevent rulers from committing injustice.' The verbs evapw, peT:aeeVGpw, pEraeevGplC w 'to form' or 'to transform' proceed from this same meaning, in the physical or moral sense: avo~povEC; eVGPovvra i ro tC; TijC; 1 VX17~ XE(!~eGLV, of ~i TWV TOLWV~E ~a~povec; -rOtC; rij~ Gocpl17~ 'fools are formed by the acquisitions of chance; but men who know [what] these acquisitions [are worth], by those of wisdom'; ij ~L~axi] pEraevapoi rov av()ewnov 'instruction transforms man'; aVclYXfJ ... ax~paTa pEraeev8plCEa()aL 'it is indeed necessary that the Gx~pa-ra change in form (in order to pass from angular to round).' Democritus also uses the adjective lnLeeVGpWc;, whose meaning can now be corrected; it is not "courant, qui se repand" (Bailly) or "adventitious" (Liddell-Scott), but 'possessing a form': hefj ov~iv tGpEV nEe£ ov~ev6c;, aAX lnLeevap{fJ iXclGrOLaW ij ~O~LC; 'we have no genuine knowledge of anything, but everyone gives a form to his belief' (= lacking knowledge of anything, everyone makes up his own opinion about everything). Accordingly, there is no variation, no ambiguity in the meaning that Democritus assigns to ev()poc;, and this is always 'form,' understood as the distinctive form, the characteristic arrangement of the parts in a whole. This point being established, there is no difficulty in confirming it by the total number of ancient examples. Let us first consider the word in Ionian prose. It is found once in Herodotus (5. 58), along with the verb pE-rae(!V()p{Cw, in a passage which is particularly interesting because it deals with the "form" of the letters of the alphabet: C'The Greeks borrowed the letters of their writing from the Phoenicians";) pEra ~i xeovov neofJalvoVToc; opa rfj cpwvfj pE-rEfJaAov xat TOV eo()pov rwv yeappclrwv 'as time passed, at the same time that they changed their language, the Cadmeans also changed the form (ev()poc;) of the characters'; Ot naea}.afJovu;c; ("Iwvec;) ~L~axfj naea TcOV t/Jowlxwv -ra yeclppa-ra, peT,aeeo()plaanEC; acpewv d}'lya lxeEWV-ro 'the lonians borrowed letters from the Phoenicians through instruction, and used them after having transformed (pEraeev()plGav-rec;) them a little.' It is not chance that Herodotus used ev() poc; for the 'form' of letters at almost the same period that Leucippus, as we have just seen, was defining this word by using the very same example. This is proof of an even more ancient tradition that applied ev()poc; to the configuration of the signs of writing. The word remained in use among the authors of the Corpus Hippocraticum, and with the same sense. One of them prescribes, for the treatment of clubfoot, the use of a small leaden boot, 'in the form of the ancient sandals of Chios' {olov at

ra

PROBLEMS IN GENERAL LINGUISTICS

xiaoeerplMe~

ev(Jp,ovElXov).5 Fromev(Jp,o~, come the compoundsop,oeevO'p,o~, 'of the same form,' opoeevO'p,{YJ 'resemblance' (Hip. 915h, 916b), EV(!eVO'p,o~ 'of a beautiful Jorm, elegant,' etc. H we now turn to the lyric poets, it is even earlier, as early as the seventh century, that we see the appearance of evO'p,o~. It is taken, like O'xii/la or ieono~, as defining the individual and distinctive 'form' of the human character. Archilochus counsels, "do not boast of your victories in public and do not collapse at home in order to weep over your defeats; rejoice at reasons for joy and do not exacerbate yourself unduly over evils; ylyvwQ";"e t5' 0 lo~ evO'p,o~ av(Jewnov~ lXet 'learn to know the dispositions which men have' " (2. 400. Bergk). In Anacreon, the evap,ol are also particular 'forms' of mood or character: eyw de p,tO'EW naVia~ oaot m~oAlOv~ lxovO't eva/love; uat xaA,enov~ (fr. 74. 2), and Theognis counts ev(Jp,o~ among man's distinctive traits: p~noi' enatv~an~ net'll a'll eldfi~ livdea O'aq;YJvw~ oey~v uat ev(Jpov uat ieonov oVitV' lXet 'never praise a man before knowing clearly his feelings, his disposition (evO'p,ot;), his character' (964). Let us add here Theocritus: •AViovoa~ eo(Jp,o~ WViO~ 'the attitude of Autonoe was the same' (26. 23). Among the tragedians, ev(Jpo~ and the verbs derived from it invariably maintain the same sense as in all the texts cited: EV ietywvot~ ev(Jp,oir; 'in a triangular Jorm,' in a fragment of Aeschylus (fr. 78 N2); VYJAEW~ wd' eeev(JtO'p,at 'a pitiless fate has made my presentJorm (= condition), (Prom. 243); noeov p,EieeeV(JptCe '(Xerxes, in his madness,) wanted to transform a strait' (Pers. 747); po'Voeev(Jpot t5op,ot 'a dwelling arranged for one person' (Suppl. 961).6 The use of ev(J plCW in Sophocles is very instructive (Antig. 318): to the guard whom Creon has commanded to be quiet because his voice makes him suffer and who asks him, "Is it in your ears or in your soul that my voice makes you suffer?" Creon replies, i{ de eV(J/l{~Et~ i~'V EP,~V AvnYJ'V onov'why do you picture the location of my grief?' Here is the exact sense of ev(Jp,lCw 'to give a form,' and the scholiast correctly renders ev(Jp,{Cet'V by O'XYJP,atlCEtV, dWionoVv 'to picture, to localize.' Euripides speaks of the evOp,o~ of a garment,ofits distinctive 'form' (ev(Jp,o~ nEnA,w'V,Heraci. 130); of the 'modality' of a murder (ieonOe; uat evOpde; q;ovov, El. 772); of the 'distinctive mark' of mourning (ev()p,d~ uauw'V, Suppl. 94); he usesEvev()p,W~ 'in a suitable fashion,' for the arrangement of a bed (Cycl. 563) and aeev()p,o~ for a 'disproportioIJate' passion (Hipp. 529). This meaning of evOp,oe; persists in the Attic prose of the fifth century. Xenophon (Mem. 3. 10. 10) makes ev(Jp,o~ ('proportion') the quality of a fine cuirass, which he qualifies by evev(Jpoe; 'of a beautiful form.' In Plato one finds, among others, the ev(Jp,o~ the 'balanced state' between opulence and poverty (Laws 728e), and expressions like ev()p,ICetV i